Step optimization method of complex industrial process, medium and system

Abstract

The invention provides a step optimization method, medium and system for a complex industrial process, and belongs to the technical field of big data modeling, and the method comprises the steps: building a multi-dimensional data collection system, installing a sensor and a monitoring device at a key node of a target industrial process, collecting process parameter data, equipment state data and product quality data in a parallel multi-thread mode to form an original data set, and collecting process flow connection relation data; performing time domain decomposition operation on the original data set, decomposing each data into a long-term stable numerical component and a short-term fluctuation numerical component, and respectively obtaining a process parameter stable component, a process parameter fluctuation component, an equipment state stable component and an equipment state fluctuation component; according to the technical scheme of the invention, the problem that the product cannot stably operate in the optimal state can be solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of big data modeling, and in particular, relates to a method, medium and system for optimizing steps in a complex industrial process. Background Art

[0002] At present, complex industrial processes such as large chemical plants, steel mills, and energy companies are facing severe challenges in product quality control. These process flows are usually composed of multiple key processes in series, and the process parameters are highly coupled with the equipment status, which has a complex impact on the quality of the final product. For example, the reactor process in a chemical production line is affected by process parameters such as temperature, pressure, and flow rate, as well as equipment status such as motor load and bearing vibration. There is a complex nonlinear relationship between these parameters, and they will attenuate and couple with the transmission of the process flow, which brings great uncertainty to the stability of product quality.

[0003] In response to the above problems, existing technologies usually adopt empirical parameter debugging or simulation optimization methods based on physical models. The empirical debugging method relies on the rich experience of process experts to explore the optimal parameter settings through trial and error. It is not only inefficient, but also difficult to ensure the reliability and repeatability of the final results. Modeling methods based on physical mechanisms, such as reaction kinetics models, heat and mass transfer models, etc., although they can more accurately describe the overall process, are also subject to the limitations of model assumptions and are difficult to fully reflect the complex situations in actual production. On the other hand, these optimization methods are usually limited to a certain process, lack global considerations, and are difficult to achieve systematic optimization of the entire process.

[0004] Therefore, there is an urgent need for a more systematic and intelligent process optimization method that can not only fully consider the coupling effects of process parameters and equipment status, but also have the ability of simulation verification and adaptive adjustment to ensure that product quality is stable under optimal conditions. This is the core technical problem that the present invention attempts to solve. Summary of the invention

[0005] In view of this, the present invention provides a method, medium and system for optimizing the steps of a complex industrial process, which can solve the problem that the product cannot operate stably under the optimal state.

[0006] The present invention is achieved in that:

[0007] A first aspect of the present invention provides a method for optimizing steps of a complex industrial process, which comprises the following steps:

[0008] S10. Establish a multi-dimensional data acquisition system, install sensors and monitoring devices at key nodes of the target industrial process, and collect process parameter data, equipment status data, and product quality data in a parallel multi-threaded manner to form an original data set, and collect process flow connection relationship data;

[0009] S20. Perform time-domain decomposition operations on the original data set, decompose each item of data into a long-term stable numerical component and a short-term fluctuation numerical component, and respectively obtain a process parameter stable component, a process parameter fluctuation component, an equipment status stable component, and an equipment status fluctuation component;

[0010] S30. Establish a mapping relationship equation set between the process parameter fluctuation component, the equipment status fluctuation component, and the product quality data, and generate a process parameter influence coefficient matrix and an equipment status influence coefficient matrix;

[0011] S40. Based on the process parameter influence coefficient matrix and the equipment status influence coefficient matrix, calculate a process operation stability matrix and an equipment operation stability matrix, and perform numerical combination operations on the process operation stability matrix, the equipment operation stability matrix, the process parameter stable component, and the equipment status stable component to construct a process mathematical model;

[0012] S50. Use digital mapping technology to construct the simulation environment of the target industrial process, input the process mathematical model into the simulation environment for numerical verification, and calculate the process parameter correction amount and the equipment parameter correction amount;

[0013] S60. Based on the process parameter correction amount and the equipment parameter correction amount, adjust the parameters of the pilot line segment of the target industrial process, collect pilot parameter data, and obtain an optimization effect value by calculating the deviation value between the pilot parameter data and the product quality data;

[0014] S70. Calculate the operation error value based on the deviation value between the pilot parameter data and the product quality data, calculate the operation correction ratio value, the operation correction integral value, and the operation correction differential value based on the operation error value, construct an automatic adjustment control unit based on the optimization effect value, generate an operation correction instruction using a preset instruction correction equation according to the operation error value, the operation correction ratio value, the operation correction integral value, and the operation correction differential value, and implement dynamic correction of the process parameters of the target industrial process based on the operation correction instruction;

[0015] S80. Calculate the deviation between the dynamic correction result of the process parameters and the product quality data, and generate a process parameter optimization guidance plan.

[0016] Among them, the process parameter data includes production temperature data, production pressure data, production speed data, and production time data;

[0017] The device status data includes device operation status data, device load rate data, device vibration data, and device temperature data;

[0018] The product quality data includes product dimension data, product performance data, product appearance data, and product defect data.

[0019] Furthermore, the mapping relation equation set includes a parameter influence mapping equation, a fluctuation attenuation mapping equation, a feature separation mapping equation, and a coefficient calculation mapping equation.

[0020] Furthermore, the parameter influence mapping equation generates quality influence feature data by performing a linear correlation operation on the product quality data, the process parameter fluctuation component, and the device status fluctuation component.

[0021] Furthermore, the fluctuation attenuation mapping equation generates fluctuation attenuation data by performing an iterative operation on the attenuation influence of the process parameter fluctuation component and the device status fluctuation component in the process flow connection relation data.

[0022] Furthermore, the feature separation mapping equation generates feature principal component data by performing a principal component decomposition operation on the quality influence feature data and the fluctuation attenuation data.

[0023] Furthermore, the coefficient calculation mapping equation generates the process parameter influence coefficient matrix and the device status influence coefficient matrix by performing a numerical correspondence operation on the feature principal component data and the quality influence feature data.

[0024] Furthermore, the target industrial process pilot line segment selects a single complete production line in the process flow for optimization verification;

[0025] The process operation specification document includes process parameter set values, process parameter adjustment ranges, process parameter dynamic correction strategies, and device operation parameter limits.

[0026] The second aspect of the present invention provides a computer-readable storage medium, in which program instructions are stored. When the program instructions run on a computer, they are used to execute the step optimization method of a complex industrial process described above.

[0027] The step optimization system of a complex industrial process provided by the third aspect of the present invention further includes the computer-readable storage medium.

[0028] The present invention proposes a method for optimizing the steps of a complex industrial process, which makes full use of data-driven modeling and analysis techniques. Through a series of algorithms such as time-domain decomposition, parameter mapping, and stability evaluation, a comprehensive mathematical model describing the relationship between process, equipment, and quality is established. On this basis, digital simulation and adaptive control technologies are also adopted to achieve intelligent optimization of the production line parameters. Compared with the prior art, this method has the following outstanding advantages:

[0029] 1. The data-driven modeling method can be closer to the actual production situation. Compared with the methods relying on empirical debugging or physical mechanism models, the method of the present invention uses the process parameters, equipment status, and product quality data collected on-site to obtain the quantitative relationship between influencing factors through mathematical modeling, which is more suitable for the complexity of industrial processes.

[0030] 2. The systematic optimization strategy realizes the collaborative improvement of the entire process flow. The prior art is usually limited to the optimization of a single process, while the method of the present invention starts from a global perspective, comprehensively considers the interaction of various process parameters and equipment status, and proposes an optimization scheme applicable to the entire production line.

[0031] 3. The simulation verification and adaptive adjustment mechanism ensure the reliability of the optimization scheme. Before actual pilot optimization, the method of the present invention first conducts sufficient parameter verification in the digital twin model, effectively avoiding production losses that may be caused by blind adjustment. At the same time, a closed-loop automatic control system is also constructed, which can automatically fine-tune the process parameters according to the real-time production situation to ensure the continuous stability of product quality.

[0032] 4. Strong generalizability and universality. The method of the present invention is not targeted at a specific industry but has wide applicability. As long as the necessary process, equipment, and quality data can be collected for the target industrial process, this optimization method can be applied to systematically improve quality.

[0033] Generally speaking, the complex industrial process optimization method proposed by the present invention gives full play to advanced technical means such as data analysis, simulation, and adaptive control, realizes intelligent optimization of the entire production process, and provides a systematic solution for quality control of complex industrial processes. Brief Description of the Drawings

[0034] Figure 1 It is a flow chart of a method for optimizing the steps of a complex industrial process. Detailed Embodiment

[0035] To make the purpose, technical solution, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0036] Such asFigure 1 As shown in the figure, it is a flowchart of a method for optimizing the steps of a complex industrial process provided by the present invention, including the following steps:

[0037] S10. Establish a multi-dimensional data acquisition system, install sensors and monitoring devices at key nodes of the target industrial process, and collect process parameter data, equipment status data, and product quality data in a parallel multi-threaded manner to form an original data set, and collect process flow connection relationship data;

[0038] S20. Perform time-domain decomposition operation on the original data set, decompose each data into a long-term stable numerical component and a short-term fluctuation numerical component, and respectively obtain the process parameter stable component, process parameter fluctuation component, equipment status stable component, and equipment status fluctuation component;

[0039] S30. Establish a mapping relationship equation set between the process parameter fluctuation component, equipment status fluctuation component, and product quality data, and generate a process parameter influence coefficient matrix and an equipment status influence coefficient matrix;

[0040] S40. Based on the process parameter influence coefficient matrix and the equipment status influence coefficient matrix, calculate the process operation stability matrix and the equipment operation stability matrix, and perform numerical combination operations on the process operation stability matrix, the equipment operation stability matrix, the process parameter stable component, and the equipment status stable component to construct a process mathematical model;

[0041] S50. Use digital mapping technology to construct a simulation environment for the target industrial process, input the process mathematical model into the simulation environment for numerical verification, and calculate the process parameter correction amount and the equipment parameter correction amount;

[0042] S60. Based on the process parameter correction amount and the equipment parameter correction amount, adjust the parameters of the pilot line segment of the target industrial process, collect the pilot parameter data, and obtain the optimization effect value by calculating the deviation value between the pilot parameter data and the product quality data;

[0043] S70. Calculate the operation error value based on the deviation value between the pilot parameter data and the product quality data, calculate the operation correction ratio value, operation correction integral value, and operation correction differential value based on the operation error value, construct an automatic adjustment control unit based on the optimization effect value, generate an operation correction instruction using a preset instruction correction equation according to the operation error value, operation correction ratio value, operation correction integral value, and operation correction differential value, and realize dynamic correction of the process parameters of the target industrial process based on the operation correction instruction;

[0044] S80. Calculate the deviation between the dynamic correction result of the process parameters and the product quality data, and generate a process parameter optimization guidance plan.

[0045] Among them, in the above technical solution, the process parameter data includes production temperature data, production pressure data, production speed data, and production time data;

[0046] The equipment status data includes equipment operation status data, equipment load rate data, equipment vibration data, and equipment temperature data;

[0047] The product quality data includes product dimension data, product performance data, product appearance data, and product defect data.

[0048] Furthermore, in the above technical solution, the mapping relation equation set includes a parameter influence mapping equation, a fluctuation attenuation mapping equation, a feature separation mapping equation, and a coefficient calculation mapping equation.

[0049] Furthermore, in the above technical solution, the parameter influence mapping equation generates quality influence feature data through performing a linear correlation operation on the product quality data, the process parameter fluctuation component, and the equipment status fluctuation component.

[0050] Furthermore, in the above technical solution, the fluctuation attenuation mapping equation generates fluctuation attenuation data through performing an iterative operation on the attenuation influence of the process parameter fluctuation component and the equipment status fluctuation component in the process flow connection relation data.

[0051] Furthermore, in the above technical solution, the feature separation mapping equation generates feature principal component data through performing a principal component decomposition operation on the quality influence feature data and the fluctuation attenuation data.

[0052] Furthermore, in the above technical solution, the coefficient calculation mapping equation generates a process parameter influence coefficient matrix and an equipment status influence coefficient matrix through performing a numerical correspondence operation on the feature principal component data and the quality influence feature data.

[0053] Furthermore, in the above technical solution, the target industrial process pilot line segment selects a single complete production line in the process flow for optimization verification;

[0054] The process operation specification document includes process parameter set values, process parameter adjustment ranges, process parameter dynamic correction strategies, and equipment operation parameter limits.

[0055] It also includes S90. Form a process operation specification document based on the process parameter optimization guidance plan to complete the overall optimization of the target industrial process.

[0056] The following describes the specific implementation manners of the above steps in detail:

[0057] The specific implementation of step S10 is as follows: Establish a multi-dimensional data acquisition system. First, install various sensors and monitoring devices at key nodes of the target industrial process, such as raw material preparation, pretreatment, main reaction, post-treatment, and finished products. The sensors include temperature sensors, pressure sensors, flow rate sensors, etc., and the monitoring devices include motor current detection, vibration monitoring, load monitoring, etc. The process parameter data (temperature, pressure, speed, etc.), equipment status data (load, vibration, temperature, etc.), and product quality data (dimensions, performance, appearance, defects, etc.) are synchronously collected in a parallel multi-threaded manner to form an original data set. At the same time, the connection relationships between the processes of the process flow are also collected and recorded, laying a foundation for the subsequent establishment of a process mathematical model.

[0058] The specific implementation of step S20 is as follows: Perform time-domain decomposition processing on the original data set. The empirical mode decomposition (EMD) algorithm is used to perform multi-scale decomposition on various types of collected data. The EMD method can adaptively decompose the non-linear and non-stationary industrial process signals into intrinsic mode functions (IMFs) with different frequencies and a residual term. During the decomposition process, first, the extreme value points are identified to divide the signal into a set of maximum value points and a set of minimum value points, and then the upper and lower envelope lines are constructed using cubic spline interpolation to obtain the first-layer intrinsic mode function. Then, an iterative method is used to continuously extract the high-frequency fluctuation component and the low-frequency stable component until the residual meets the convergence condition.

[0059] Through this time-domain decomposition operation, the process parameter data and the equipment status data can be decomposed into a long-term stable component and a short-term fluctuation component, which are respectively denoted as the process parameter stable component, the process parameter fluctuation component, the equipment status stable component, and the equipment status fluctuation component. The purpose of this step is to lay a foundation for the subsequent establishment of an impact mapping relationship.

[0060] The specific implementation of step S30 is as follows: Establish a mapping relationship between parameter fluctuations and product quality. First, by performing a linear correlation analysis on the product quality data, the process parameter fluctuation component, and the equipment status fluctuation component, the influence degree of each parameter fluctuation on the product quality is determined to generate quality influence characteristic data. Secondly, considering that the parameter fluctuations will attenuate and propagate in the process flow, an exponential decay function is used to model the propagation law of parameter fluctuations to obtain fluctuation attenuation data.

[0061] Then, the principal component analysis (PCA) method is used to extract features and reduce the dimension of the quality impact feature data and the fluctuation attenuation data, obtaining the principal component feature data. Finally, the process parameter influence coefficient matrix and the equipment state influence coefficient matrix are calculated by the least squares method. The above steps establish a mapping relationship equation system between parameter fluctuations and product quality, providing a basis for subsequent optimization.

[0062] The specific implementation manner of step S40 is: constructing a process mathematical model. First, based on the process parameter influence coefficient matrix and the equipment state influence coefficient matrix obtained in step S30, the process operation stability matrix and the equipment operation stability matrix are calculated to reflect the overall stability of the system.

[0063] Secondly, the process parameter stable component, the equipment state stable component are numerically combined with the above two stability matrices to establish a comprehensive process mathematical model. This model uses the method of linear weighting, that is where the weight coefficient w 1 , w 2 , w 3 By optimizing the objective function J = ∑(Y measured -Y model ) 2 to solve, satisfying the constraint condition of w 1 + w 2 + w 3 = 1. The random error ω is also considered in the model to improve the adaptability and robustness of the model.

[0064] The purpose of this step is to establish a mathematical model that can describe the overall characteristics of the target industrial process, comprehensively reflecting the impacts of process parameters, equipment states and their interactions on product quality, providing a basis for subsequent simulation optimization.

[0065] The specific implementation manner of step S50 is: constructing a simulation environment based on digital mapping technology. First, digital mapping technologies such as virtual reality (VR) or digital twin are used to establish a simulation environment highly consistent with the actual production line. This simulation environment includes not only the geometric model of the process flow, but also the mathematical models of physical processes such as heat and momentum transfer.

[0066] Next, the process mathematical model established in step S40 is input into this simulation environment, and the correction amounts of process parameters and equipment parameters are obtained through numerical simulation calculations. During the simulation process, different process conditions and equipment states can also be set specifically to observe their impacts on product quality, providing decision-making support for the optimization of the actual production line.

[0067] The purpose of this step is to optimize process parameters and equipment parameters in a computer simulation environment, avoiding production losses caused by blind attempts on the actual production line.

[0068] The specific implementation of step S60 is as follows: Conduct parameter optimization verification on the actual production line. First, according to the process parameter correction amount and equipment parameter correction amount obtained in step S50, adjust the parameters of the pilot line segment of the target industrial process. The selection of the pilot line segment is usually a single complete production line in the whole process to ensure that the optimization results are representative.

[0069] Next, collect the adjusted parameter data on the pilot line segment and measure the quality indicators of the products. By calculating the deviation value between the pilot parameter data and the product quality data, a numerical evaluation of the optimization effect can be obtained. This evaluation result is the basis for the subsequent automatic adjustment control strategy.

[0070] The purpose of this step is to verify the effectiveness of the simulation optimization in the actual production environment and lay a foundation for further automatic optimization control.

[0071] The specific implementation of step S70 is as follows: Construct an automatic adjustment control unit. First, use the pilot parameter data and product quality data obtained in step S60 to calculate the operation error value of the actual production process. Then, based on this operation error value, use the Proportional-Integral-Derivative (PID) control algorithm to calculate the correction ratio value, correction integral value, and correction differential value of the process parameters.

[0072] Next, based on the optimization effect value and the above PID control parameters, construct an automatic adjustment control unit. This control unit takes the operation error value, correction ratio value, correction integral value, and correction differential value as inputs and generates specific dynamic correction instructions for process parameters through a preset instruction correction equation.

[0073] The purpose of this step is to establish a closed-loop automatic control mechanism to achieve continuous optimization of the target industrial process and ensure the stable operation of product quality in the best state.

[0074] The specific implementation of step S80 is as follows: Form a process parameter optimization guidance plan. Calculate the deviation between the dynamic correction result of the process parameters implemented in step S70 and the product quality data to obtain a quantitative evaluation of the optimization effect. Based on this, formulate a targeted process parameter optimization guidance plan to provide a basis for subsequent overall optimization.

[0075] The optimization guidance plan includes, but is not limited to: the optimal setting value range of process parameters, the dynamic adjustment range of process parameters, the sorting of the sensitivity of each parameter to product quality, etc. Through these guidelines, the focus of process optimization can be pointed out, providing a clear optimization direction for operators.

[0076] The specific implementation method of step S90 is: complete the overall optimization and form an operating procedure. Further refine the process parameter optimization guidance plan obtained in step S80 to formulate a detailed process operating procedure document. This document includes the set values of process parameters, adjustment intervals, dynamic correction strategies, and the limit requirements of equipment operation parameters, etc.

[0077] The compilation of the operating procedure document can not only guide production personnel to operate according to the optimization plan, but also serve as the basis for process management and equipment management. By strictly implementing this operating procedure, it can ensure that the target industrial process operates stably in the optimal state and continuously meets the product quality requirements.

[0078] Based on the content of the present invention, the equations for each calculation process are described in detail:

[0079] 1. Time-domain decomposition operation:

[0080] The mathematical expression of the time-domain decomposition operation is specifically as follows:

[0081]

[0082] In the formula, x(t) is the original collected signal, including process parameter data and equipment status data; c i (t) is the i-th intrinsic mode function, representing the fluctuation component; r n (t) is the residual term, representing the stable component; n is the decomposition layer number, with a value of 3 to 5.

[0083] The decomposition steps include:

[0084] (1) Obtain the extreme points:

[0085] E max ={(t i ,x i )|x i >x i-1 ,x i >x i+1 ;

[0086] E min ={(t i ,x i )|x i <x i-1 ,x i <x i+1};

[0087] In the formula, E max is the set of maximum points; E min is the set of minimum points; t i is the time point; x i is the signal value.

[0088] (2) Construct the envelope:

[0089]

[0090] In the formula, s up (t) is the upper envelope; s low (t) is the lower envelope; both are obtained by cubic spline interpolation.

[0091] (3) Judge the IMF condition:

[0092]

[0093] In the formula, SD is the standard deviation; θ is the threshold, with a value range of 0.2 - 0.3; k is the number of iterations; T is the data length.

[0094] 2. Parameter influence mapping:

[0095] The parameter influence mapping equation is specifically expressed as follows:

[0096]

[0097] In the formula, Q ij is the quality influence characteristic data; ΔP ik is the process parameter fluctuation component; ΔE jk is the equipment state fluctuation component; α k , β k are the weight coefficients; γ ij is the random error, with a range of 0.01 - 0.1; m is the number of fluctuation components.

[0098] Weight coefficient calculation:

[0099]

[0100]

[0101] In the formula, Corr(·) represents the correlation coefficient calculation function.

[0102] 3. Fluctuation attenuation mapping:

[0103] The fluctuation attenuation mapping equation is specifically expressed as follows:

[0104]

[0105] In the formula, Dij is the fluctuation attenuation data; k is the process sequence number; λ k , η k are the attenuation coefficients; μ k , ν k are the attenuation exponents; ξ ij is the systematic error, with a range of 0.05 to 0.15.

[0106] Attenuation parameter calculation:

[0107]

[0108] In the formula, λ 0 , η 0 are the initial attenuation coefficients; a p , a e is the attenuation rate, obtained by fitting historical data.

[0109] 4. Feature principal component decomposition:

[0110] The specific expression of the feature principal component decomposition is as follows:

[0111] F = VQ;

[0112] In the formula, F is the feature principal component matrix; V is the feature vector matrix; Q is the quality influence feature matrix.

[0113] Feature vector calculation:

[0114] CV = λV;

[0115]

[0116] In the formula, C is the covariance matrix; λ is the eigenvalue; N is the number of samples.

[0117] 5. Influence coefficient calculation:

[0118] The specific expression of the influence coefficient matrix calculation is as follows:

[0119] M p = F T Q p / ||F|| 2 + ∈ p ;

[0120] M e = F T Q e / ||F|| 2 + ∈ e ;

[0121] In the formula, M p , M e are the influence coefficient matrices of process parameters and equipment status respectively; ∈p , ∈ e is the fitting error, with a range of 0.02 - 0.08.

[0122] 6. Stability evaluation:

[0123] The calculation of the stability matrix is specifically expressed as follows:

[0124]

[0125] In the formula, S p , S e are the process and equipment stability matrices respectively; δ p , δ e is the correction term, with a range of 0.1 - 0.3.

[0126] 7. Process model construction:

[0127] The process mathematical model is specifically expressed as follows:

[0128]

[0129] In the formula, Y is the model output; P s , E s are the stable components; w 1 , w 2 , w 3 are the weight coefficients; is the Kronecker product; ω is the model error, with a range of 0.1 - 0.2.

[0130] 8. Parameter correction:

[0131] The calculation of the parameter correction amount is specifically expressed as follows:

[0132]

[0133] In the formula, ΔP is the parameter correction amount; e is the running error; K p , K i , K d are the PID control parameters.

[0134] Explanation of the equation principle:

[0135] 1. The EMD method is used to achieve multi-scale decomposition, which can adaptively process non-linear and non-stationary signals;

[0136] 2. The linear superposition principle describes the influence of parameters and considers random errors to improve the robustness of the model;

[0137] 3. The exponential decay model describes the characteristics of wave propagation and conforms to the law of energy dissipation;

[0138] 4. Use the PCA method to reduce the dimension and extract features, reducing redundant information;

[0139] 5. Use the least squares method to determine the influence coefficient, ensuring the fitting accuracy;

[0140] 6. Use matrix norm to evaluate the stability, reflecting the overall characteristics of the system;

[0141] 7. Use the Kronecker product to describe parameter interaction, reflecting the coupling effect;

[0142] 8. Use PID control to achieve closed-loop optimization, ensuring the dynamic performance.

[0143] Now, the derivation process of each equation will be described in detail.

[0144] The time-domain decomposition operation equation uses the empirical mode decomposition method, and its mathematical expression is This equation can decompose the complex signals in the industrial process into wave components of different scales. Among them, the original signal x(t) includes temperature signals (such as the temperature of the reaction kettle, heating temperature, cooling temperature, etc.), pressure signals (such as system pressure, local pressure, pressure difference, etc.), speed signals (such as material conveying speed, stirring speed, flow rate, etc.), and equipment signals (such as motor current, bearing vibration, equipment load, etc.) in the industrial field. The decomposition process first identifies the extreme points E max ={(t i , x i )|x i >x i-1 , x i >x i+1} and E min ={(t i , x i )|x i <x i-1 , x i <x i+1}, and then uses cubic spline interpolation to construct the envelope According to the time scale, the components are divided into wave components and stable components, where the time scale thresholds of different parameters are determined according to the process characteristics, such as temperature change of 300s, pressure fluctuation of 60s, speed fluctuation of 30s, and equipment status of 600s.

[0145] The parameter influence mapping equation describes the influence relationship between process parameters and equipment status fluctuations on product quality, and the expression is The wave component of process parameters ΔP ik includes temperature fluctuation (±2°C), pressure fluctuation (±0.1 MPa), speed fluctuation (±5%), and time fluctuation (±30 s); the wave component of equipment status ΔE jkincluding load fluctuation (±10%), vibration fluctuation (±0.5 mm / s), temperature rise fluctuation (±5 °C), and current fluctuation (±15%); quality impact characteristic Q ij It includes product dimensions (length, width, thickness), product performance (strength, hardness, density), product appearance (surface quality, color difference, gloss), and product defects (cracks, bubbles, impurities), etc.

[0146] Fluctuation attenuation mapping equation Describes the transfer and attenuation law of parameter fluctuations in the process flow. The process sequence number k corresponds to processes such as raw material preparation (k = 1), pretreatment (k = 2), main reaction (k = 3), post-treatment (k = 4), and finished product (k = 5) in actual production. The attenuation coefficient is calculated through and where different parameters have different initial attenuation coefficients and attenuation rates, such as temperature attenuation (λ 0 = 1.0, a p = 0.2), pressure attenuation (λ 0 = 0.8, a p = 0.3), speed attenuation (λ 0 = 0.9, a p = 0.25), and equipment attenuation (η 0 = 0.95, a e = 0.15).

[0147] Process mathematical model Comprehensively considers the stable components of process parameters, the stable components of equipment states, and their interactions. The stable components include the stable values of process parameters (temperature set value, pressure set value, speed set value) and the stable values of equipment states (equipment load stable value, equipment vibration reference value, equipment temperature reference value). The weight coefficients are solved through the optimization objective function J = ∑(Y measured - Y model ) 2 while satisfying the constraint condition that the sum of weights is 1.

[0148] The industrial process optimization method of the present invention has significant technical advantages and practical values compared with the traditional manual experience adjustment method. From a technical perspective, this method establishes a complete data-driven decision-making mechanism. It obtains process parameters, equipment status, and product quality data in real time through a multi-dimensional data acquisition system, and uses advanced mathematical models to accurately model and analyze complex industrial systems. The core lies in using time-domain decomposition technology to decompose complex signals into stable components and fluctuation components, accurately describing the parameter influence mechanism through a system of mapping relationship equations, achieving an accurate mathematical description of the process, and avoiding the subjectivity and uncertainty of traditional empirical models. This data- and model-based method can accurately quantify the influence relationships between parameters, predict the impact of parameter adjustments on product quality, and provide a reliable theoretical basis for optimization decisions.

[0149] From the perspective of control strategies, the present invention breaks through the limitations of traditional fixed-parameter adjustments and constructs a closed-loop optimization system with dynamic adaptive capabilities. By establishing an automatic adjustment control unit, the system can calculate the running error and correction amount in real time, automatically generate and execute running correction instructions, and achieve dynamic optimization adjustments of process parameters. This intelligent control mechanism can not only adapt to changes in working conditions and keep process parameters in the optimal state at all times, but also greatly reduce manual intervention, improve production efficiency, and product quality stability. Especially in a complex and changeable industrial environment, this method shows extremely strong adaptability and reliability.

[0150] From the perspective of systematics and standardization, the present invention adopts an overall optimization idea and realizes multi-parameter collaborative optimization through matrix operations and numerical mapping techniques. The system not only considers the influence of individual parameters, but also pays more attention to the interaction between parameters, and finds the global optimal solution through scientific trade-off analysis. The finally formed standardized process operation procedures specify in detail the parameter setting values, adjustment ranges, and correction strategies, providing clear operation guidance for industrial production. This standardized method greatly improves the popularization of technology and overcomes the problems that traditional empirical methods are difficult to standardize and are prone to experience loss due to personnel changes.

[0151] From the perspective of actual application effects, the present invention significantly improves the intelligent level of industrial production by establishing a full-process automation system from data acquisition, parameter optimization to instruction execution.

[0152] The second aspect of the present invention provides a computer-readable storage medium, in which program instructions are stored. When the program instructions run on a computer, they are used to execute the step optimization method of a complex industrial process as described above.

[0153] The third aspect of the present invention provides a step optimization system for a complex industrial process, which further includes the computer-readable storage medium.

[0154] Specifically, the principle of the present invention is as follows:

[0155] 1. Multi-dimensional data acquisition and time-domain decomposition: For complex industrial processes, various sensors and monitoring devices are first deployed at key nodes to collect process parameter data (such as temperature, pressure, speed, etc.), equipment status data (such as load, vibration, temperature rise, etc.), and product quality data (such as size, performance, appearance, etc.). These raw data often have complex non-linear and non-stationary characteristics. Therefore, the present invention uses the empirical mode decomposition (EMD) algorithm to decompose various types of data into long-term stable components and short-term fluctuation components. This adaptive time-domain decomposition method can better capture the dynamic change laws of parameters and states in the process flow.

[0156] 2. Mapping relationship between parameter fluctuations and quality: Based on the above decomposition results, the present invention establishes a mapping relationship between process parameter fluctuations, equipment status fluctuations, and product quality. First, the influence degree of each parameter fluctuation on quality is determined through correlation analysis; secondly, considering the attenuation characteristics of parameter fluctuations in the process, an exponential decay function is used for modeling; finally, methods such as principal component analysis are used to extract key features, and a linear regression model is established to obtain the influence coefficient matrix. This series of mapping relationship equations provides a necessary quantitative basis for subsequent process optimization.

[0157] 3. Construction of process mathematical model: On the basis of establishing the parameter-quality mapping relationship, the present invention further constructs a mathematical model describing the entire process flow. This model comprehensively considers the stable components of process parameters, the stable components of equipment status, and their interaction, and obtains the final product quality prediction value through weighted linear combination. At the same time, modeling and evaluation are also carried out for process operation stability and equipment operation stability to reflect the overall characteristics of the system. Such a process mathematical model lays a solid foundation for subsequent digital simulation and adaptive optimization.

[0158] 4. Digital simulation and adaptive optimization: Based on the above process mathematical model, the present invention uses technologies such as virtual reality or digital twin to construct a digital simulation environment highly consistent with the actual production line. In this simulation environment, process parameters and equipment status can be fully optimized and verified, and the best parameter correction amount can be calculated.

[0159] Subsequently, the present invention constructs an automatic adjustment control unit to achieve real-time optimization feedback of production line parameters. This control unit dynamically adjusts process parameters using the PID control algorithm according to the operation error in the actual production process to ensure that the product quality continuously meets the requirements. Through the mechanism of simulation verification and adaptive adjustment, the solution of the present invention can effectively avoid the risks that may be brought by blind adjustment and improve the reliability of the optimization solution.

[0160] A specific embodiment 1 of the present invention is provided below. The specific implementation methods of each step in this embodiment 1 are described in detail as follows:

[0161] The specific implementation method of step S10 is as follows: First, various sensors and monitoring devices are installed at key nodes of the target industrial process, such as raw material preparation, pretreatment, main reaction, post-treatment, and finished products. The sensors used include temperature sensors, pressure sensors, flow rate sensors, etc., and the monitoring devices include motor current detection, vibration monitoring, load monitoring, etc. In order to achieve high-speed data acquisition and parallel processing, the parallel multi-threaded method is used to synchronously collect process parameter data (temperature T, pressure P, speed V, etc.), equipment status data (load L, vibration R, temperature T e etc.) and product quality data (dimensions D, performance F, appearance A, defects D f etc.) to form the original data set X. At the same time, the connection relationship G between each process of the process flow is also collected and recorded, laying a foundation for establishing a process mathematical model later. The purpose of this step is to construct a multi-dimensional data acquisition system to provide necessary original data support for subsequent data analysis and optimization.

[0162] The specific implementation method of step S20 is as follows: Perform time-domain decomposition processing on the original data set X obtained in step S10. The empirical mode decomposition (EMD) algorithm is used to perform adaptive multi-scale decomposition on various types of collected data. The EMD method can decompose the non-linear and non-stationary industrial process signal into intrinsic mode functions (IMFs) c i (t) and a residual term r n (t). The mathematical expression is:

[0163]

[0164] where x(t) is the original collected signal, c i (t) is the i-th intrinsic mode function (representing the fluctuation component), r n (t) is the remaining term (representing the stable component), and n is the number of decomposition layers, with the value range of 3 to 5.

[0165] The decomposition process includes: First, by identifying the extreme points E max ={(t i ,x i )|x i >x i-1 ,x i >x i+1} and E min ={(t i ,x i )|x i <x i-1 ,x i <xi+1 , construct the envelope using cubic spline interpolation Next, continuously extract the high-frequency fluctuation components and low-frequency stable components in an iterative manner until the residual satisfies the convergence condition SD = where the value range of θ is 0.2 to 0.3.

[0166] Through this time-domain decomposition operation, the process parameter data and equipment status data can be decomposed into long-term stable components P s , E s and short-term fluctuation components ΔP, ΔE, laying a foundation for establishing the influence mapping relationship subsequently. The purpose of this step is to extract multi-scale features from the original data, which is conducive to analyzing the internal relationship between parameter fluctuations and quality subsequently.

[0167] The specific implementation method of step S30 is: establish the mapping relationship between parameter fluctuations and product quality. First, through linear correlation analysis of the product quality data Q with the process parameter fluctuation component ΔP and the equipment status fluctuation component ΔE, determine the influence degree of each parameter fluctuation on the product quality, and generate quality influence characteristic data Q ij , and its mathematical expression is:

[0168]

[0169] where α k , β k are weight coefficients, obtained by calculating the correlation coefficient Corr:

[0170]

[0171] γ ij is a random error, and its value range is 0.01 to 0.1.

[0172] Secondly, considering that parameter fluctuations will attenuate and propagate in the process flow, use the exponential decay function to model the propagation law of parameter fluctuations and obtain the fluctuation attenuation data D ij , and its mathematical expression is:

[0173]

[0174] where λ 0 , η 0 are the initial attenuation coefficients, a p , a e is the attenuation rate, ξ ij is the system error, and its value range is 0.05 to 0.15.

[0175] Then, the principal component analysis (PCA) method is used to extract features and reduce the dimension of the quality impact feature data Q and the fluctuation attenuation data D, obtaining the principal component feature matrix F, where F = VQ, and V is the feature vector matrix, is the covariance matrix, and λV = CV is the eigenvalue-eigenvector relationship.

[0176] Finally, the process parameter influence coefficient matrix M p and the equipment state influence coefficient matrix M e are calculated by the least squares method. The mathematical expressions are:

[0177] M p = F T Q p / ||F|| 2 + ∈ p ;

[0178] M e = F T Q e / ||F|| 2 + ∈ e ;

[0179] where ∈ p , ∈ e is the fitting error, and the value range is 0.02 to 0.08.

[0180] The above steps establish a mapping relationship equation set between parameter fluctuations and product quality, providing a basis for subsequent optimization.

[0181] The specific implementation of step S40 is: constructing a process mathematical model. First, based on the process parameter influence coefficient matrix M p and the equipment state influence coefficient matrix M e obtained in step S30, the process operation stability matrix S p and the equipment operation stability matrix S e are calculated. The mathematical expressions are:

[0182]

[0183] where δ p , δ e is the correction term, and the value range is 0.1 to 0.3. These two matrices reflect the overall stability of the system.

[0184] Secondly, the process parameter stable component P s , the equipment state stable component E s are numerically combined with the above two stability matrices to establish a comprehensive process mathematical model Y. The mathematical expression is:

[0185]

[0186] Among them, w 1 , w 2 , w 3 are weight coefficients, and are solved by optimizing the objective function J = ∑(Y measured - Y model ), satisfying the constraint condition of w 2 + w 1 + w 2 + w 3 = 1; ω is the model error, and the value range is 0.1 to 0.2. This model can describe the influence of process parameters, equipment status and their interactions on product quality, and provide a basis for subsequent simulation optimization.

[0187] The specific implementation of step S50 is: constructing a simulation environment based on digital mapping technology. First, technologies such as virtual reality (VR) or digital twin are used to establish a simulation environment highly consistent with the actual production line, including geometric models and physical process models.

[0188] Next, the process mathematical model established in step S40 is input into this simulation environment, and the correction amount ΔP of process parameters and the correction amount ΔE of equipment parameters are obtained through numerical simulation calculations. During the simulation process, different process conditions and equipment status can also be set specifically to observe their influence on product quality.

[0189] The purpose of this step is to optimize process parameters and equipment parameters in a computer simulation environment, avoiding production losses caused by blind attempts on the actual production line.

[0190] The specific implementation of step S60 is: conducting parameter optimization verification on the actual production line. First, according to the correction amounts of process parameters and equipment parameters obtained in step S50, parameter adjustments are made to the pilot segment of the target industrial process. The selection of the pilot segment is usually a single complete production line in the entire process to ensure that the optimization results are representative.

[0191] Then, the adjusted parameter data is collected on the pilot segment, and the quality indicators of the product are measured. By calculating the deviation value between the pilot parameter data and the product quality data, a numerical evaluation of the optimization effect J = ∑(Y measured - Y model ) 2 can be obtained. This evaluation result is the basis for the subsequent automatic control strategy.

[0192] The purpose of this step is to verify the effectiveness of the simulation optimization in the actual production environment and lay a foundation for further automated optimization control.

[0193] The specific implementation of step S70 is as follows: Construct an automatic adjustment control unit. First, using the pilot parameter data and product quality data obtained in step S60, calculate the running error value e of the actual production process, where e = Y measured - Y model . Then, based on this running error value, use the proportional-integral-derivative (PID) control algorithm to calculate the correction ratio value K p of the process parameter, the correction integral value K i , and the correction differential value K d . The mathematical expression is:

[0194]

[0195] Next, based on the optimization effect value and the above PID control parameters, construct an automatic adjustment control unit. This control unit takes the running error value e, the correction ratio value K p , the correction integral value K i , and the correction differential value K d as inputs, and generates specific dynamic correction instructions for process parameters through a preset instruction correction equation.

[0196] The purpose of this step is to establish a closed-loop automatic control mechanism, realize the continuous optimization of the target industrial process, and ensure the stable operation of product quality in the best state.

[0197] The specific implementation of step S80 is as follows: Form a process parameter optimization guidance plan. Calculate the deviation between the dynamic correction result of the process parameter implemented in step S70 and the product quality data to obtain a quantitative evaluation of the optimization effect. Based on this, formulate a targeted process parameter optimization guidance plan to provide a basis for subsequent overall optimization.

[0198] This optimization guidance plan includes but is not limited to: the optimal setting value range of process parameters (temperature T s ∈ [150°C, 180°C], pressure P s ∈ [0.5 MPa, 0.8 MPa], speed V s ∈ [80 m / min, 100 m / min], etc.), the dynamic adjustment range of process parameters (ΔT ∈ ±2°C, ΔP ∈ ±0.1 MPa, ΔV ∈ ±5%, etc.), the sensitivity ranking of each parameter to product quality, etc. Through these guidelines, the focus of process optimization can be pointed out, providing a clear optimization direction for operators.

[0199] The specific implementation of step S90 is as follows: Complete the overall optimization and form an operating procedure. Further refine the process parameter optimization guidance plan obtained in step S80 to formulate a detailed process operating procedure document. This document includes:

[0200] 1) Set values of process parameters, such as temperature T s = 165 °C, pressure P s = 0.65 MPa, speed V s = 90 m / min, etc.;

[0201] 2) Adjustment ranges of process parameters, such as temperature T ∈ [163 °C, 167 °C], pressure P ∈ [0.6 MPa, 0.7 MPa], speed V ∈ [87 m / min, 93 m / min], etc.;

[0202] 3) Dynamic correction strategies for process parameters, such as using the PID control algorithm in step S70 for closed-loop regulation;

[0203] 4) Limit requirements for equipment operating parameters, such as load L < 90%, vibration R < 0.6 mm / s, temperature rise ΔT e < 10 °C, etc.

[0204] The compilation of the operating procedure document can not only guide production personnel to operate according to the optimized plan, but also serve as a basis for process management and equipment management. By strictly implementing this operating procedure, it can ensure that the target industrial process operates stably in the optimal state and continuously meets the product quality requirements.

[0205] The following provides a specific embodiment 2 of the present invention for optimizing the process parameters of a chemical production line. The specific methods for each step in this embodiment 2 are described in detail as follows:

[0206] A certain chemical plant produces a high-performance polymer product. The entire production process includes multiple key processes such as raw material preparation, preheating reaction, main reaction, and post-treatment. Since the precise control of process parameters such as reaction temperature, pressure, and stirring speed has a great impact on product quality, and in addition, changes in equipment status (such as the load of the reaction kettle motor and bearing vibration) will also interfere with process stability, it has always been a major technical problem for this plant.

[0207] Adopt the complex industrial process optimization method proposed by the present invention. The specific implementation steps are as follows:

[0208] Step S10: Establish a multi-dimensional data acquisition system. Temperature sensors, pressure sensors, rotational speed sensors, etc. are installed at key equipment such as reaction kettles, heat exchangers, and mixers. At the same time, motor current detectors, vibration monitoring equipment, etc. are also equipped to achieve comprehensive acquisition of process parameters and equipment status. The main data collected includes: reaction temperature T, reaction pressure P, stirring rotational speed V, motor current I, bearing vibration R, product size D, product strength F, product appearance A, etc. The data acquisition frequency is set to 1 Hz to ensure that the dynamic changes in the process can be accurately captured. At the same time, the process flow connection relationship G between each process is also recorded.

[0209] Step S20: Perform time-domain decomposition on the collected data. First, by identifying the extreme points, construct the upper and lower envelopes of parameters such as temperature, pressure, and rotational speed using the cubic spline interpolation method. Then, use the EMD algorithm to adaptively decompose these signals, dividing them into long-term stable components T s , P s , V s and short-term fluctuation components ΔT, ΔP, ΔV. After decomposition, it is found that the temperature fluctuation is within the range of ±2°C, the pressure fluctuation is within the range of ±0.1 MPa, and the rotational speed fluctuation is within the range of ±5%. For the equipment status data, the characteristics of current fluctuation ΔI ∈ ±15% and vibration fluctuation ΔR ∈ ±0.5 mm / s are also identified.

[0210] Step S30: Establish the mapping relationship between parameter fluctuations and product quality. First, through the correlation analysis of quality indicators such as product size D, strength F, and appearance A with process parameter fluctuations and equipment status fluctuations, it is found that temperature fluctuation ΔT and motor load fluctuation ΔI have a greater impact on product size, pressure fluctuation ΔP and bearing vibration fluctuation ΔR mainly affect product strength, and rotational speed fluctuation ΔV and temperature rise fluctuation ΔT e will lead to a decline in product appearance quality. Based on this, the following linear regression models are established:

[0211] D = 0.8ΔT + 0.6ΔI + ∈ D ;

[0212] F = 0.7ΔP + 0.5ΔR + ∈ F ;

[0213] A = 0.9ΔV + 0.7ΔT e + ∈ A ;

[0214] where ∈ D , ∈ F , ∈ A is the random error term.

[0215] At the same time, considering the attenuation characteristics of parameter fluctuations in the process flow, an exponential decay function is established to describe its propagation law:

[0216] ΔT = ΔT 0 e -0.2k , ΔP = ΔP 0 e -0.3k , ΔV = ΔV 0 e -0.25k ;

[0217] ΔI = ΔI 0 e -0.15k , ΔR = ΔR 0e -0.18k , ΔT e = ΔT e0 e -0.12k ;

[0218] Among them, k is the process serial number, which reflects the transfer attenuation characteristics of parameters in the process.

[0219] Step S40: Construct a process mathematical model. Based on the above mapping relationship, a mathematical model describing the performance of the entire production line is further established:

[0220]

[0221] Among them, S T , S P , S V are the stability indexes of temperature, pressure, and rotational speed respectively, and S I , S R , S T are the stability indexes of current, vibration, and temperature rise, and ω is the model error. This mathematical model can quantitatively describe the comprehensive influence of process parameters, equipment states, and their interactions on product quality.

[0222] Step S50: Optimize and verify based on the simulation environment. Using digital twin technology, a simulation environment highly consistent with the actual production line is established, including 3D geometric models of main equipment such as reactors, heat exchangers, and mixers, as well as corresponding heat and momentum transfer models. The above process mathematical model is imported into this simulation environment, and the optimal correction amounts of process parameters are obtained through numerical calculations:

[0223] ΔT = -1.5°C, ΔP = +0.08 MPa, ΔV = +3%;

[0224] ΔI = -10%, ΔR = -0.3 mm / s, ΔT e = +4°C;;

[0225] Step S60: Verify the parameter optimization on the pilot production line. According to the simulation optimization results, on-site pilot verification is carried out on a complete polymer production line. The adjusted process parameters and product quality data are collected, and the optimization effects are obtained through calculation: the qualified rate of product dimensions is increased from 92% to 96%; the qualified rate of product strength is increased from 89% to 94%; the qualified rate of product appearance is increased from 85% to 92%.

[0226] Step S70: Construct an automatic regulation control system. Based on the optimization effect data of the pilot production line, a PID feedback control strategy is established to monitor the product quality deviation in real time and automatically adjust the process parameters:

[0227]

[0228] Among them, T set is the temperature set value, K p , K i , K d are the PID control parameters and can be optimized online according to the actual production situation.

[0229] Steps S80 and S90: Form an optimized guidance plan for process parameters and compile an operating procedure. Based on the above optimization practice, the following optimized suggestions for process parameters are formulated:

[0230] 1) The temperature set value T s = 165 °C, and the control range is T ∈ [163 °C, 167 °C];

[0231] 2) The pressure set value P s = 0.65 MPa, and the control range is P ∈ [0.6 MPa, 0.7 MPa];

[0232] 3) The rotational speed set value V s = 90 rpm, and the control range is V ∈ [87 rpm, 93 rpm];

[0233] 4) The limit value of the motor current I max = 85 A, the limit value of vibration R max = 0.6 mm / s, and the limit value of temperature rise ΔT e,max = 10 °C;

[0234] At the same time, a detailed process operating procedure is also compiled, which stipulates the dynamic adjustment strategy of the above parameters and provides a clear quality control basis for production personnel.

[0235] Compared with the traditional manual experience debugging method, the solution of the present invention has the following advantages:

[0236] 1. The data-driven modeling method is closer to the actual situation and can accurately reflect the internal relationship between process - equipment - quality. Compared with the debugging method relying on expert experience, the solution of the present invention uses a large amount of data collected on site and obtains a quantitative influence relationship through statistical analysis and mathematical modeling, which is more in line with the actual situation of complex process.

[0237] 2. The systematic optimization strategy improves the overall effect. The traditional method is usually limited to the parameter adjustment of a single process, while the solution of the present invention starts from the overall situation, comprehensively considers the coupling influence of each process link and equipment state, and gives an optimization solution applicable to the entire production line. The pilot verification data shows that the qualified rate of each quality index of the product has increased by an average of 5 percentage points, and the overall performance has been significantly improved.

[0238] 3. The simulation verification and adaptive adjustment mechanism ensure the reliability of the optimization scheme. Before actual pilot optimization, sufficient parameter verification is carried out in the digital twin model, effectively avoiding production losses caused by blind adjustment. At the same time, the present invention also constructs a closed-loop automatic control system, which can automatically fine-tune process parameters according to the real-time production situation to ensure the continuous stability of product quality. This intelligent feature greatly improves the executability of the scheme.

[0239] Generally speaking, the solution of the present invention gives full play to cutting-edge technical means such as data analysis, simulation optimization, and adaptive control, realizes the systematic optimization of process parameters of complex chemical production lines, and provides an effective solution for quality control in this industry. Compared with traditional empirical debugging, it not only has better optimization effects, but also has stronger replicability and universality.

[0240] The following provides a specific Embodiment 3 of the present invention for optimizing the manufacturing process of aircraft components. The specific methods of each step in this Embodiment 3 are described in detail as follows:

[0241] An aviation enterprise produces an aircraft wing leading edge component with a complex structure. The manufacturing process involves multiple key processes such as metal cutting, heat treatment, and surface treatment. Since machining parameters (such as rotational speed, feed rate, cutting depth, etc.), heat treatment temperature, surface treatment current, etc. have a great impact on the component performance and appearance quality, and there is also a coupling relationship with equipment status (such as tool wear, fixture stability, power supply fluctuation, etc.), it has always been a technical difficulty for this enterprise.

[0242] Adopt the complex industrial process optimization method proposed by the present invention. The specific implementation steps are as follows:

[0243] Step S10: Establish a multi-dimensional data acquisition system. Install rotational speed sensors, feed sensors, temperature sensors, current sensors, vibration sensors, etc. on key equipment such as numerically controlled machine tools, heat treatment furnaces, and electroplating lines to achieve comprehensive monitoring of process parameters and equipment status. The main data collected include: cutting rotational speed V, feed rate F, cutting depth D, heat treatment temperature T, electroplating current I, component size D p , component strength F p , component appearance A p , etc. The data acquisition frequency is set to 5Hz, and the process flow connection relationship G of each process is recorded.

[0244] Step S20: Perform time-domain decomposition on the collected data. First, by identifying extreme points, the upper and lower envelopes of machining parameters and heat treatment parameters are constructed using the cubic spline interpolation method. Then, the EMD algorithm is used to adaptively decompose these signals and divide them into long-term stable components V s , F s , D s , Ts , I s and short - term fluctuation components ΔV, ΔF, ΔD, ΔT, ΔI. After decomposition, it is found that the rotational speed fluctuation is within ±3, the feed fluctuation is within ±5%, the cutting depth fluctuation is within ±0.05 mm, the heat treatment temperature fluctuation is within ±3 °C, and the electroplating current fluctuation is within ±10%. For the equipment status data, the characteristics of tool wear fluctuation ΔW ∈ ±15%, fixture vibration fluctuation ΔR ∈ ±0.4 mm / s, and power supply voltage fluctuation ΔU ∈ ±4% are also identified.

[0245] Step S30: Establish the mapping relationship between parameter fluctuations and component quality. By performing a correlation analysis on component quality indicators such as component size D p , strength F p , appearance A p etc. and process parameter fluctuations and equipment status fluctuations, it is found that the cutting speed fluctuation ΔV and the tool wear fluctuation ΔW have a greater impact on the component size, the feed fluctuation ΔF and the fixture vibration fluctuation ΔR mainly affect the component strength, and the cutting depth fluctuation ΔD and the power supply fluctuation ΔU will lead to a decline in the component appearance quality. Based on this, the following linear regression model is established:

[0246]

[0247] where, is the random error term.

[0248] At the same time, an exponential decay function is also established to describe the propagation law of parameter fluctuations in the process flow:

[0249] ΔV = ΔV 0 e -0.3k , ΔF = ΔF 0 e -0.35k , ΔD = ΔD 0 e -0.4k ;

[0250] ΔT = ΔT 0 e -0.25k , ΔI = ΔI 0 e -0.2k ;

[0251] ΔW = ΔW 0 e -0.18k , ΔR = ΔR 0 e -0.22k , ΔU = ΔU 0 e -0.15k ;

[0252] Step S40: Construct a process mathematical model. Based on the above mapping relationship, a mathematical model describing the performance of the entire manufacturing line is further established:

[0253]

[0254] Among them, S V , S F , S D , S T , S I are the stability indexes of rotational speed, feed, cutting depth, temperature, and current respectively. S W , S R , S U are the stability indexes of tool wear, fixture vibration, and power supply fluctuation. ω is the model error. This model can quantitatively describe the comprehensive influence of machining parameters, heat treatment parameters, equipment status, and their interactions on the quality of components.

[0255] Step S50: Perform optimization verification based on the simulation environment. Using digital twin technology, a simulation environment highly consistent with the actual manufacturing line is established, including 3D geometric models of main equipment such as CNC machine tools, heat treatment furnaces, and electroplating lines, as well as corresponding mechanical and heat transfer models. The above process mathematical model is imported into this simulation environment, and the optimal correction amounts of process parameters are obtained through numerical calculation:

[0256] ΔV = -2%, ΔF = -3%, ΔD = -0.03 mm;

[0257] ΔT = -2 °C, ΔI = -8%;

[0258] ΔW = -10%, ΔR = -0.2 mm / s, ΔU = -3%;

[0259] Step S60: Conduct parameter optimization verification on the pilot production line. According to the simulation optimization results, on-site pilot verification is carried out on a complete production line for wing leading edge components. The adjusted process parameters and component quality data are collected, and the optimization effects are obtained through calculation: the qualified rate of component dimensions is increased from the original 90% to 95%; the qualified rate of component strength is increased from the original 88% to 93%; the qualified rate of component appearance is increased from the original 86% to 92%.

[0260] Step S70: Construct an automatic adjustment control system. Based on the optimization effect data of the pilot production line, a PID feedback control strategy is established to monitor the component quality deviation in real time and automatically adjust the process parameters:

[0261]

[0262] Among them, V set , F set , D set are the set values of machining parameters, and K p , K i , K d are the PID control parameters, which can be optimized online according to the actual production situation.

[0263] Steps S80 and S90: Form an optimized process parameter guidance plan and compile an operating procedure. Based on the above optimization practices, the following optimized process parameter suggestions are formulated:

[0264] 1) Cutting speed set value V s = 2500 rpm, control range V ∈ [2450 rpm, 2550 rpm];

[0265] 2) Feed rate set value F s = 0.2 mm / s, control range F ∈ [0.18 mm / s, 0.22 mm / s];

[0266] 3) Depth of cut set value D s = 0.8 mm, control range D ∈ [0.75 mm, 0.85 mm];

[0267] 4) Heat treatment temperature set value T s = 950 °C, control range T ∈ [945 °C, 955 °C];

[0268] 5) Electroplating current set value I s = 25 A, control range I ∈ [23 A, 27 A];

[0269] 6) Upper limit of tool wear W max = 20%, upper limit of fixture vibration R max = 0.6 mm / s, power supply voltage deviation ΔU max = ±5%;

[0270] Meanwhile, a detailed process operating procedure is also compiled, which stipulates the dynamic adjustment strategy of the above parameters, providing a clear quality control basis for production personnel.

[0271] Compared with the traditional manual experience debugging method, the solution of the present invention has the following advantages:

[0272] 1. The system optimization of the whole process improves the overall quality level. The traditional method usually only adjusts the parameters for a single process, ignoring the coupling relationship between process links, and it is difficult to achieve the overall improvement of quality from a global perspective. In contrast, the solution of the present invention comprehensively considers the influencing factors of the whole process of machining, heat treatment, and surface treatment, and gives an optimized solution accordingly. The pilot verification data shows that the pass rate of each quality index of the components has increased by an average of 5 percentage points.

[0273] 2. The data-driven modeling method is more accurate and reliable. Traditional empirical debugging relies on the subjective judgment of process experts and it is difficult to quantify the quantitative relationships between parameters. However, the present invention utilizes a large amount of data collected on-site and adopts statistical analysis and mathematical modeling methods, which can more scientifically reflect the internal laws of process - equipment - quality.

[0274] 3. The simulation verification and adaptive adjustment mechanism ensure the feasibility of the solution. Before actual pilot optimization, full parameter verification is first carried out in the digital twin model, effectively avoiding production losses that may be caused by blind adjustment. At the same time, the present invention also constructs a closed-loop automatic control system, which can automatically fine-tune process parameters according to the real-time production situation to ensure that the component quality continuously meets the requirements. This intelligent feature greatly improves the operability of the solution.

[0275] It should be noted that the variables involved in the present invention are explained in detail in Table 1 as follows:

[0276] Table 1 Variable Explanation Table

[0277]

[0278] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention.

Claims

1. A method for optimizing steps in a complex industrial process, characterized in that: The following steps are involved: S10. Establish a multi-dimensional data collection system, install sensors and monitoring equipment at key nodes of the target industrial process, use parallel multi-threading to collect process parameter data, equipment status data, product quality data, form an original data set, and collect process connection relationship data; S20, performing time domain decomposition operation on the original data set, decomposing each data into a long-term stable numerical component and a short-term fluctuating numerical component, and obtaining a process parameter stable component, a process parameter fluctuating component, an equipment state stable component, and an equipment state fluctuating component respectively; S30, establishing a mapping relationship equation group between the process parameter fluctuation component, the equipment state fluctuation component and the product quality data, and generating a process parameter influence coefficient matrix and an equipment state influence coefficient matrix; S40, based on the process parameter influence coefficient matrix and the equipment state influence coefficient matrix, calculate the process operation stability matrix and the equipment operation stability matrix, and perform numerical combination operation on the process operation stability matrix, the equipment operation stability matrix, the process parameter stable component, and the equipment state stable component to construct a process mathematical model; S50, using digital mapping technology to construct the target industrial process simulation environment, inputting the process mathematical model into the simulation environment for numerical verification, and calculating the process parameter correction amount and the equipment parameter correction amount; S60, adjusting the parameters of the target industrial process pilot line segment based on the process parameter correction amount and the equipment parameter correction amount, collecting pilot parameter data, and obtaining an optimization effect value by calculating the deviation between the pilot parameter data and the product quality data; S70, calculating an operation error value based on the deviation value between the pilot parameter data and the product quality data, calculating an operation correction proportional value, an operation correction integral value, and an operation correction differential value based on the operation error value, constructing an automatic adjustment control unit according to the optimization effect value, generating an operation correction instruction using a preset instruction correction equation according to the operation error value, the operation correction proportional value, the operation correction integral value, and the operation correction differential value, and realizing dynamic correction of the target industrial process parameter based on the operation correction instruction; S80, performing deviation calculation on the process parameter dynamic correction result and the product quality data to generate a process parameter optimization guidance plan.

2. A method for optimizing steps of a complex industrial process according to claim 1, characterized in that: The process parameter data include production temperature data, production pressure data, production speed data, and production time data; The equipment status data includes equipment operation status data, equipment load rate data, equipment vibration data, and equipment temperature data; The product quality data includes product size data, product performance data, product appearance data, and product defect data.

3. A method for optimizing steps of a complex industrial process according to claim 2, characterized in that: The mapping relationship equation group includes a parameter influence mapping equation, a fluctuation attenuation mapping equation, a feature separation mapping equation, and a coefficient calculation mapping equation.

4. A method for optimizing steps of a complex industrial process according to claim 3, characterized in that: The parameter impact mapping equation generates quality impact characteristic data by performing linear correlation operations on the product quality data, the process parameter fluctuation component, and the equipment state fluctuation component.

5. A method for optimizing steps of a complex industrial process according to claim 4, characterized in that: The fluctuation attenuation mapping equation generates fluctuation attenuation data by iteratively calculating the attenuation effects of the process parameter fluctuation component and the equipment status fluctuation component in the process flow connection relationship data.

6. A method for optimizing steps of a complex industrial process according to claim 5, characterized in that: The feature separation mapping equation generates feature principal component data by performing principal component decomposition operation on the quality-affecting feature data and the fluctuation attenuation data.

7. A method for optimizing steps of a complex industrial process according to claim 6, characterized in that: The coefficient calculation mapping equation generates the process parameter influence coefficient matrix and the equipment state influence coefficient matrix by performing numerical correspondence operations on the characteristic principal component data and the quality influence characteristic data.

8. A method for optimizing steps of a complex industrial process according to claim 7, characterized in that: The target industrial process pilot line segment selects a single complete production line in the process flow for optimization verification; The process operation procedure document includes process parameter setting values, process parameter adjustment ranges, process parameter dynamic correction strategies, and equipment operation parameter limits.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program instructions, and when the program instructions are run in a computer, they are used to execute the step optimization method for a complex industrial process according to any one of claims 1 to 8.

10. A system for optimizing steps in a complex industrial process, characterized in that: Also included is the computer-readable storage medium of claim 9.

Images