Blood sample collector production scheduling optimization system based on cloud platform

Abstract

The invention relates to the technical field of production process management, in particular to a blood sample collector production scheduling optimization system based on a cloud platform. Comprising a situation quantification unit used for collecting multi-source heterogeneous data, calculating to obtain an efficiency pressure index and a quality risk probability, and generating a production comprehensive conflict index according to the efficiency pressure index and the quality risk probability; the dynamic decision-making unit is used for comparing the production comprehensive conflict index with a preset first decision-making threshold value and a preset second decision-making threshold value so as to generate a real-time optimal instruction; the risk prediction unit is used for predicting the fluctuation quantity of future key physical parameters based on the historical comprehensive state vector sequence and future event information, and generating a feedforward pre-compensation instruction according to the fluctuation quantity; and the fusion control unit is used for fusing the real-time optimal instruction and the feed-forward pre-compensation instruction to generate a composite control instruction. According to the system, dependence on artificial experience or static rules is converted into accurate and unified numerical evaluation, and an objective basis is provided for subsequent intelligent decision making.

Description

Technical Field

[0001] This invention relates to the field of production process management technology, specifically a cloud-based blood sample collector production scheduling optimization system. Background Technology

[0002] As a sophisticated medical device, the production process of blood sample collectors involves multiple stages, including injection molding, automated assembly, and the critical application of heparin sodium anticoagulant coating. This process places extremely high demands on both production efficiency and product quality. Among these, critical processes such as coating application are irreversible, and any quality issues will result in the scrapping of the entire batch of products. In the production scheduling of blood sample collectors, there is always a core contradiction between pursuing production efficiency and ensuring product quality. On the one hand, enterprises need to maximize equipment utilization and meet the urgent delivery deadlines of orders, which brings huge efficiency pressure. On the other hand, it is necessary to strictly control the environmental parameters of the clean room and ensure the batch stability of raw materials in order to avoid potential quality risks. Existing production scheduling systems typically rely on static rules or human experience for decision-making, making it difficult to assess the combined situation of current efficiency pressures and quality risks in real time and quantitatively. When both efficiency pressures and quality risks increase simultaneously, traditional scheduling methods cannot dynamically adjust production strategies and lack the ability to predict and proactively avoid future risks, such as fluctuations in process parameters caused by material changes. This passive and static scheduling model makes the production system extremely vulnerable to complex dynamic changes, making it difficult to effectively balance the contradiction between efficiency and quality. This poses a significant risk of batch scrapping due to quality problems, affecting production efficiency and order delivery capabilities.

[0003] The information disclosed in the background section above is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention discloses a cloud-based blood sample collector production scheduling optimization system. Specifically, the technical solution of this invention is as follows: A cloud-based blood sample collector production scheduling and optimization system includes: The contextual quantification unit is used to collect multi-source heterogeneous data, calculate the efficiency pressure index and quality risk probability based on the multi-source heterogeneous data, and generate a comprehensive production conflict index based on the efficiency pressure index and quality risk probability. The dynamic decision-making unit is used to compare the comprehensive production conflict index with the preset first decision threshold and second decision threshold to generate the real-time optimal instruction; The risk prediction unit is used to predict the fluctuation of key physical parameters in the future based on historical comprehensive state vector sequences and future event information, and generate feedforward pre-compensation instructions accordingly. The fusion control unit is used to fuse real-time optimal commands and feedforward pre-compensation commands to generate composite control commands.

[0005] Preferably, the scenario quantification unit is specifically used to: obtain the overall equipment utilization rate and order urgency; and perform a weighted summation of the overall equipment utilization rate and order urgency to generate an efficiency pressure index; the overall equipment utilization rate is calculated based on the real-time collected equipment start-up time, cycle time, and the number of good products and total output; the order urgency is calculated as a function of the ratio of the remaining delivery time of the order to the standard production period.

[0006] Preferably, the context quantification unit is further specifically used to: obtain the environmental cleanliness index and the material stability score; and perform a weighted summation of the environmental cleanliness index and the material stability score to generate a quality risk probability; the environmental cleanliness index is generated by a weighted comprehensive evaluation of multiple environmental parameters; the material stability score is the result of a weighted summation and normalization of material batch inspection data and supplier historical quality ratings.

[0007] Preferably, the specific operation of the dynamic decision-making unit is as follows: When the overall production conflict index is less than or equal to the first decision threshold, the efficiency-first strategy is generated as the real-time optimal instruction. When the overall production conflict index is greater than the first decision threshold and less than or equal to the second decision threshold, a balance strategy is generated as the real-time optimal instruction. When the overall production conflict index exceeds the second decision threshold, a quality-first strategy is generated as the real-time optimal instruction.

[0008] Preferably, the risk prediction unit uses a long short-term memory network model to learn a comprehensive state vector sequence that includes a historical comprehensive conflict index sequence, key physical parameters, and supply chain parameters, so as to output a comprehensive state vector evolution sequence for multiple future time steps, and extract the fluctuation of future key physical parameters from the comprehensive state vector evolution sequence.

[0009] Preferably, the generation process of the feedforward pre-compensation command is as follows: Based on the fluctuations of key physical parameters in the future, and using a pre-set process response surface model, the compensation amount of the process parameters is calculated to generate feedforward pre-compensation instructions.

[0010] Preferably, the fusion control unit is used to quantize the real-time optimal command into a dimensionless real-time adjustment factor.

[0011] Preferably, the fusion control unit is further configured to apply the compensation amount contained in the feedforward pre-compensation command and the real-time adjustment factor together to a preset reference process parameter to generate a composite control command.

[0012] Preferably, the comprehensive production conflict index is generated through a nonlinear weighted model, and its calculation formula is as follows: ,in, Efficiency pressure index. For the probability of quality risk, and These are the basic weighting coefficients for efficiency and quality, respectively.

[0013] Preferably, the formula for calculating the efficiency pressure index is: ,in, Efficiency pressure index. The normalized equipment utilization rate To normalize the order urgency, Assign weights to efficiency.

[0014] Preferably, the formula for calculating the probability of quality risk is: For the probability of quality risk, The environmental cleanliness index, To score material stability, Assign weights to risks.

[0015] Preferably, the formula for generating the composite control command is: It is a composite control command. As the baseline process parameters, This refers to the compensation amount included in the feedforward pre-compensation instruction.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. This system quantifies complex production site conditions into a single comprehensive production conflict index by collecting data from multiple sources, including equipment, environment, and supply chain. This transforms the assessment of the core conflict between pursuing efficiency and ensuring quality from relying on manual experience or static rules to a precise and unified numerical assessment, providing an objective basis for subsequent intelligent decision-making.

[0017] 2. This system establishes a data-driven closed-loop feedback mechanism by comparing the overall production conflict index with a preset decision threshold. This mechanism can automatically and smoothly switch between different strategies such as efficiency priority, equilibrium, and quality priority based on the intensity of production conflicts, enabling the production system to automatically select the optimal operating mode for the current working conditions.

[0018] 3. This system incorporates a risk prediction unit based on historical data learning, enabling proactive risk avoidance for critical irreversible process steps. It can anticipate potential fluctuations in key physical parameters caused by events such as material switching and generate feedforward pre-compensation instructions, thus transforming passive response into proactive prevention and significantly reducing the risk of batch scrap due to quality issues.

[0019] 4. This system integrates the control unit to combine real-time optimal instructions for responding to current operating conditions with feedforward pre-compensation instructions for mitigating future risks. This integration forms a composite control logic that combines immediacy and foresight, ensuring that the final production instructions can not only respond keenly to current changes but also cope with foreseeable future risks, thus comprehensively improving the stability and robustness of the production process. Attached Figure Description

[0020] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of the system of the present invention. Detailed Implementation Example

[0022] Please see Figure 1 A cloud-based blood sample collector production scheduling and optimization system includes: The contextual quantification unit is used to collect multi-source heterogeneous data, calculate the efficiency pressure index and quality risk probability based on the multi-source heterogeneous data, and generate a comprehensive production conflict index based on the efficiency pressure index and quality risk probability. The dynamic decision-making unit is used to compare the comprehensive production conflict index with the preset first decision threshold and second decision threshold to generate the real-time optimal instruction; The risk prediction unit is used to predict the fluctuation of key physical parameters in the future based on historical comprehensive state vector sequences and future event information, and generate feedforward pre-compensation instructions accordingly. The fusion control unit is used to fuse real-time optimal commands and feedforward pre-compensation commands to generate composite control commands. This embodiment provides a cloud-based blood sample collector production scheduling optimization system. Its core technology lies in constructing an intelligent closed loop of perception, decision-making, prediction, and control to achieve a dynamic balance between efficiency and quality and proactive risk avoidance in the production process. The system includes a context quantification unit, a dynamic decision-making unit, a risk prediction unit, and a fusion control unit; The purpose of the contextual quantification unit is to abstract and transform the complex and multi-dimensional production site conditions into a single, calculable numerical indicator, providing a unified and accurate evaluation basis for subsequent intelligent decision-making. In this embodiment, the unit collects multi-source heterogeneous data in real time through communication protocols such as industrial Ethernet and OPCUA. These data sources include: Internet of Things (IoT) sensor data from equipment such as injection molding machines and automated assembly lines, such as mold temperature and servo motor speed; environmental parameters from cleanroom environmental monitoring systems (EMS), such as temperature, humidity, pressure difference, and dust particle count; and supply chain data from enterprise resource planning (ERP) and warehouse management systems (WMS), such as material batches, supplier historical quality ratings, and order delivery cycles. After acquiring the data, the scenario quantification unit calculated the efficiency pressure index based on the multi-source heterogeneous data. and quality risk probability Subsequently, the unit was based on the efficiency pressure index. and quality risk probability A comprehensive production conflict index is generated through a nonlinear weighted model. The Production Conflict Index χ is an indicator used to characterize the core contradiction in production in real time and quantitatively, namely the intensity of the conflict between pursuing efficiency and ensuring quality. Its role is to serve as the core driving signal of the entire intelligent scheduling system. This index is based on the efficiency pressure index and quality risk probability calculated in the preceding steps. For example, this nonlinear weighted model can be expressed as a combination of Euclidean distance and exponential amplification:

[0023] in, and These are the basic weighting coefficients for efficiency and quality, respectively, satisfying... ; is the nonlinear amplification factor, a hyperparameter greater than 0, used in... and At higher levels, the conflict index is amplified exponentially, thus more sensitively reflecting high-risk situations.

[0024] The reason for choosing a nonlinear model is that linear weighted summation might mask the true conflict: when and When one indicator is extremely high and the other is extremely low, their weighted sum may still be at a moderate level, failing to accurately reflect potential extreme risks. The nonlinear model used in this embodiment can amplify this conflict, especially when both indicators rise simultaneously. This will exhibit exponential growth, thus providing a stronger early warning signal for the decision-making system; the dynamic decision-making unit aims to quickly and automatically select the optimal production strategy based on the intensity of current production conflicts, replacing the traditional scheduling methods that rely on manual experience or static rules; in this embodiment, this unit will calculate the comprehensive production conflict index in real time. Compared with the pre-set first decision threshold Second decision threshold Comparison; First decision threshold Second decision threshold These are two key threshold values ​​that divide production into three response zones: low conflict, medium conflict, and high conflict. Their setting is based on statistical analysis of historical production data. For example, they are used when historical data shows a production plan achievement rate greater than 95% and a batch scrap rate less than 0.5%. The 80th quantile of the value distribution is set as The inflection point at which the batch scrap rate begins to show a significant non-linear increase corresponds to... Value set to Through this comparison, the unit can generate the optimal real-time instruction. The risk prediction unit aims to upgrade from passive response to proactive avoidance, especially for critical process steps such as the spraying of heparin sodium anticoating coatings, which are prone to irreversibility. In this embodiment, the unit is based on a historical comprehensive state vector sequence. Using known information about future events, such as material changeover plans in an ERP system, a pre-trained Long Short-Term Memory (LSTM) network model can predict the fluctuations of key physical parameters within a specific future time step, such as the next 3 hours. For example, predicting the viscosity of heparin sodium solution. The price will increase by 5% due to batch change; based on this prediction, the unit will further generate a feedforward pre-compensation instruction, which includes the amount of process parameter adjustment required to offset this fluctuation. The purpose of the fusion control unit is to ensure that the final instructions issued by the system can both respond sensitively to the current operating conditions and cope with future risks, forming a closed-loop control logic that combines immediacy and foresight. In this embodiment, this unit fuses the real-time optimal instructions generated by the dynamic decision-making unit and the feedforward pre-compensation instructions generated by the risk prediction unit. Specifically, it quantifies the two instructions into adjustments to the baseline process parameters and calculates them together to generate the final composite control instructions issued to the Manufacturing Execution System (MES). ; This embodiment constructs a complete intelligent closed loop of perception-cognition-decision-prediction-pre-optimization through the collaborative work of the four units mentioned above; it solves the fundamental problem that traditional static scheduling systems cannot respond in real time to the dynamic contradiction between efficiency and quality in the production process; and it quantifies complex production scenarios into a calculable conflict index. Based on this, dynamic decision-making and feedforward risk avoidance are carried out. Under the premise of ensuring the quality of the core coating of the blood sample collector and irreversible process, this system can maximize production efficiency and order delivery capability, and significantly reduce the risk of batch scrap due to quality problems.

[0025] Example 2: The contextual quantification unit is specifically used to: obtain the overall equipment utilization rate and order urgency; and perform a weighted summation of the overall equipment utilization rate and order urgency to generate an efficiency pressure index; the overall equipment utilization rate is calculated based on the real-time collected equipment start-up time, cycle time, and the number of good products and total output; the order urgency is calculated as a function of the ratio of the remaining delivery time of the order to the standard production cycle. The contextual quantification unit is also specifically used for: obtaining the environmental cleanliness index and material stability score; and performing a weighted summation of the environmental cleanliness index and material stability score to generate a quality risk probability; the environmental cleanliness index is generated by a weighted comprehensive evaluation of multiple environmental parameters; the material stability score is the result of a weighted summation and normalization of material batch inspection data and supplier historical quality ratings; This embodiment, based on Embodiment 1, explains how the context quantification unit calculates the efficiency pressure index. and quality risk probability Specific limitations were imposed to ensure the objectivity, repeatability, and accuracy of the core indicator calculation process; to ensure the robustness of the model, all input variables, such as..., were... , , , All have undergone boundary checks and validity verification; for example, to prevent division by zero errors, standard production lead times... It is constrained to be greater than a very small positive number; for sensor data, obvious outliers are removed by setting reasonable thresholds, thereby avoiding drastic fluctuations in decision instructions due to data anomalies. To calculate the efficiency stress index The context quantification unit specifically performs the following steps: Obtain the overall utilization rate of equipment and order urgency Equipment overall utilization rate It is an indicator calculated according to the industry standard algorithm (OEE), which accurately reflects the actual operating efficiency at the equipment level. It is calculated by collecting real-time data on equipment uptime, the difference between actual and theoretical production cycles, and the ratio of good products to total output; order urgency. It is an indicator reflecting time pressure at the task level. Its function is to quantify the urgency of order delivery. It is calculated by a function of the ratio of the remaining delivery time of the order to the standard production cycle. For example, its calculation function can be specifically expressed as:

[0026] in, The remaining delivery time for the order. This is the standard production lead time; this formula ensures that when the remaining time is greater than or equal to the standard lead time, It is 0; as the remaining time decreases, Increment linearly to 1; This unit uses the following formula to calculate the overall utilization rate of the equipment. and order urgency A weighted summation is performed to generate an efficiency pressure index. :

[0027] in, Efficiency stress index, [dimensionless, normalized to] The result is obtained by calculation using this formula; Normalized equipment utilization rate, [dimensionless], real-time equipment data, calculated using the OEE standard algorithm; Normalized order urgency (dimensionless), calculated from order data in the ERP system using a function. Efficiency allocation weights [dimensionless hyperparameters], whose values ​​can be calibrated using historical data to optimize a specific production objective, such as maximizing average daily output; To calculate the probability of quality risk The context quantification unit also specifically performs the following steps: Obtaining the environmental cleanliness index Material stability score Environmental cleanliness index It is a comprehensive indicator for assessing the environmental quality of the core production area. Its function is to quantify the potential impact of environmental factors on product quality. It is determined by a weighted comprehensive evaluation of multiple environmental parameters collected by the EMS system, such as temperature, humidity, pressure difference, and dust particle count. Post-generation; Material stability score It is an indicator for assessing the reliability of raw material quality. Its function is to quantify the quality risks introduced by materials. It is the result of weighted summation and normalization of multiple discrete indicators such as the purity and concentration of heparin sodium and the supplier's historical quality rating. For example, its calculation can be specified as follows:

[0028] in, , , These are normalized scores representing the purity, concentration, and supplier rating of heparin sodium, respectively. These are the corresponding weighting coefficients, satisfying... and ; This unit uses the following formula to calculate the environmental cleanliness index. Material stability score Perform a weighted summation to generate the quality risk probability. :

[0029] in, : Probability of quality risk, [dimensionless, normalized to] The result is obtained by calculation using this formula; Environmental cleanliness index, [dimensionless], is obtained by EMS multi-parameter weighted evaluation; Material stability score, [dimensionless], is obtained by weighted evaluation of WMS / ERP material and supplier data; Risk allocation weights, [dimensionless hyperparameters], are determined based on process sensitivity analysis. For example, if experimental data shows that environmental cleanliness has a much greater impact on yield than material stability, then the weights should be increased accordingly. value.

[0030] Example 3:

[0031] The specific operation of the dynamic decision-making unit is as follows: When the overall production conflict index is less than or equal to the first decision threshold, the efficiency-first strategy is generated as the real-time optimal instruction. When the overall production conflict index is greater than the first decision threshold and less than or equal to the second decision threshold, a balance strategy is generated as the real-time optimal instruction. When the overall production conflict index is greater than the second decision threshold, a quality-first strategy is generated as the real-time optimal instruction. Based on Example 1, this embodiment elaborates on the specific operational logic of the dynamic decision-making unit; its underlying logic lies in establishing a data-driven closed-loop feedback mechanism, enabling the system to automatically and smoothly switch operating modes according to the intensity of production conflicts. The specific operation of the dynamic decision-making unit is designed as a piecewise function, and its decision-making logic is as follows: When the real-time acquired comprehensive production conflict index Less than or equal to the first decision threshold ,For example The system determines that the current production state is a low-conflict zone, meaning that quality risks are controllable and efficiency pressures are not significant; at this point, the system generates an efficiency-first strategy. As the optimal instruction in real time; efficiency-first strategy It is a set of parameter configurations designed to maximize output. In the specific technical context of this invention, its special meaning is to allow the equipment to operate at near-maximum performance while sacrificing a very small quality margin. For example, the instruction may specifically include: allowing the servo motor speed of the automated assembly line to be increased to more than 95% of the rated value, and reducing the product sampling frequency of non-critical processes. When the comprehensive production conflict index Greater than the first decision threshold And less than or equal to the second decision threshold ,For example The system determines that the current production state is in a balanced zone, meaning that the conflict between efficiency and quality is beginning to emerge and a trade-off needs to be made; at this point, the system generates a balancing strategy. As the real-time optimal instruction; balancing strategy It is a set of compromise parameter configurations designed to find the optimal balance between output and quality; for example, the instruction may specifically include: moderately reducing the production cycle time by 5% compared to the baseline value, while doubling the frequency of online detection of critical coating uniformity; When the comprehensive production conflict index Greater than the second decision threshold ,For example The system determines that the current production status is a high-conflict zone, meaning that the quality risk is significantly increased, and product quality must be prioritized to avoid batch scrap. At this time, the system generates a quality-priority strategy. As the best real-time instruction; quality-first strategy It is a set of conservative parameter configurations that prioritize ensuring product qualification rate at the expense of short-term output; for example, the instruction may specifically include: immediately reducing the production speed by 30%, locking the currently used batch of materials for quality review, and triggering mandatory inspection of cleanroom environmental parameters. In a preferred embodiment, to avoid simplifying complex problems into overly idealized models, the dynamic decision-making unit determines whether to enter a high-conflict zone ( Further analysis will be conducted at that time. The main source of contribution; for example, if the calculation finds that the probability of quality risk R is related to... The contribution exceeds 80%, and the increase in R is mainly due to the material stability score. If the value is too low, a quality-first strategy will be generated. It will include more targeted instructions, such as immediately suspending the use of the current batch of materials and sending a re-inspection alert to the quality control department, rather than simply taking uniform measures to reduce the production rate, thereby improving the accuracy and practicality of decision-making.

[0032] Example 4: The risk prediction unit uses a long short-term memory network model to learn a comprehensive state vector sequence that includes historical comprehensive conflict index sequence, key physical parameters and supply chain parameters, so as to output a comprehensive state vector evolution sequence for multiple future time steps, and extract the volatility of future key physical parameters from the comprehensive state vector evolution sequence. The process of generating the feedforward pre-compensation command is as follows: Based on the fluctuation of key physical parameters in the future, and using a pre-set process response surface model, the compensation amount of the process parameters is calculated to generate feedforward pre-compensation instructions. Based on Example 1, this embodiment optimizes and limits the model selection, input and output, and feedforward pre-compensation instruction generation process of the risk prediction unit, aiming to achieve accurate prediction and proactive avoidance of irreversible risks in key processes. The risk prediction unit uses a Long Short-Term Memory (LSTM) network model. The reason for choosing the LSTM model is that its special gating structure is very suitable for learning and predicting long-term dependencies in time series data, and the production process of the blood sample collector is a complex system with multiple variables and temporal correlations. To achieve accurate predictions, this LSTM model incorporates historical comprehensive conflict indices. Sequence, key physical parameters, such as the real-time viscosity of heparin sodium solution. Real-time temperature of the cleanroom And supply chain parameters, such as the comprehensive state vector sequence of supplier IDs for the current material batch. Learning; synthesizing state vector sequences It is a multi-dimensional time series data structure whose function is to provide predictive models with comprehensive information about the past and present state of the system. It is composed of various real-time data collected and integrated by the context quantization unit. Through the Through learning, the model can output a sequence of integrated state vector evolutions over multiple future time steps, for example, over the next 8 hours, with one step per hour. The system extracts fluctuations in key future physical parameters from the evolution sequence of the comprehensive state vector; for example, by comparing the viscosity predicted 3 hours later. With current viscosity Calculate the predicted volatility ; After obtaining the fluctuation amount, the process of generating the feedforward pre-compensation instruction is as follows: This unit is based on the fluctuations of key future physical parameters. The system then uses a pre-defined process response surface model to calculate the compensation amounts for process parameters, ultimately generating feedforward pre-compensation commands. The process response surface model is a mathematical model established through extensive process experiments; its function is to describe key process parameters such as nozzle pressure and drying temperature. The quantitative relationship between the process response surface model and final product quality, such as coating uniformity, is simplified to the following first-order linear approximation in this embodiment to illustrate its principle. In practical applications, the process response surface model can be extended to a nonlinear polynomial containing cross terms and higher-order terms to more accurately describe the coupling effects between parameters.

[0033]

[0034] in, Pressure compensation amount, [unit: pressure, e.g.] The result is calculated using this formula and is used as part of the feedforward instruction. Temperature compensation amount, [unit: temperature, e.g.] The result is calculated using this formula and is used as part of the feedforward instruction. : The predicted viscosity fluctuation, [the dimension is viscosity, e.g.] The output of the LSTM model; Compensation coefficient, [dimensions are respectively] and The calibration is based on process experiment (DOE) parameters. Specifically, a calibration dataset is obtained through a series of process experiments. This dataset contains multiple sets of process parameter observations under different experimental settings, such as experimental viscosity. Experimental pressure Based on the corresponding product quality results and this calibration dataset, a response surface model describing the relationship between quality and process parameters is fitted using regression analysis methods such as the least squares method, and the compensation coefficients are obtained from it. and .

[0035] Example 5: The fusion control unit is used to quantize the real-time optimal command into a dimensionless real-time adjustment factor; The fusion control unit is also used to apply the compensation amount contained in the feedforward pre-compensation command and the real-time adjustment factor together to the preset reference process parameters to generate composite control commands. Based on Example 1, this embodiment provides a specific design for how the fusion control unit integrates instructions from two different sources to generate the final composite control instruction. This design constructs a comprehensive control system that can quickly respond to current emergencies and calmly cope with foreseeable future risks. The fusion control unit is used to process real-time optimal commands, i.e., the aforementioned Quantified as a dimensionless real-time adjustment factor Real-time adjustment factor It is a conflict index with the current situation The directly correlated adjustment coefficient serves to transform the macro-level strategy output by the upper-level decision-making unit, such as prioritizing efficiency, into a precise multiplier factor that can be used for mathematical calculations. To clarify its implementation, this quantification relationship follows pre-established quantification mapping rules, for example: When the strategy prioritizes efficiency hour, =1.05, which means a 5% increase in load on the baseline process; When the strategy is a balance strategy hour, =1.0 indicates that the baseline process parameters are maintained; When the strategy is quality first hour, =0.8 indicates that the process parameters will be reduced by 20% to ensure stability; Based on the above quantization results, the fusion control unit is also used to incorporate the compensation amount included in the feedforward pre-compensation command, such as... With real-time adjustment factor Together they act on the preset baseline process parameters To generate composite control commands; based on nozzle pressure For example, its generation formula is as follows:

[0036] This formula is the core of the entire closed-loop control logic. It integrates two different types of adjustment commands through linear superposition: the product term representing real-time feedback adjustment and the compensation term representing feedforward predictive adjustment. It also ensures the consistency of physical dimensions. The final pressure = reference pressure × dimensionless factor + pressure compensation amount. in, The final execution pressure / composite control command [with pressure as the unit] is calculated by this formula and sent to the MES system; Reference process pressure, [dimensions are pressure, e.g.] Based on the process center value preset according to the standard operating procedure (SOP); Real-time adjustment factor, [dimensionless], real-time optimal instruction. It is quantized based on mapping rules; Feedforward pressure compensation amount, [dimension is pressure, e.g.] The calculation results of the risk prediction unit.

[0037] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention; any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0038] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A cloud-based blood sample collector production scheduling and optimization system, characterized in that, include: The contextual quantification unit is used to collect multi-source heterogeneous data, calculate the efficiency pressure index and quality risk probability based on the multi-source heterogeneous data, and generate a comprehensive production conflict index based on the efficiency pressure index and quality risk probability. The dynamic decision-making unit is used to compare the comprehensive production conflict index with the preset first decision threshold and second decision threshold to generate the real-time optimal instruction; The risk prediction unit is used to predict the fluctuation of key physical parameters in the future based on historical comprehensive state vector sequences and future event information, and generate feedforward pre-compensation instructions accordingly. The fusion control unit is used to fuse real-time optimal commands and feedforward pre-compensation commands to generate composite control commands.

2. The cloud-based blood sample collector production scheduling optimization system according to claim 1, characterized in that, The scenario quantification unit is specifically used to: obtain the overall equipment utilization rate and order urgency; and perform a weighted summation of the overall equipment utilization rate and order urgency to generate an efficiency pressure index; the overall equipment utilization rate is calculated based on the real-time collected equipment start-up time, cycle time, and the number of good products and total output; the order urgency is calculated as a function of the ratio of the remaining delivery time of the order to the standard production period.

3. The cloud-based blood sample collector production scheduling optimization system according to claim 1, characterized in that, The context quantification unit is also specifically used to: obtain the environmental cleanliness index and material stability score; and to perform a weighted summation of the environmental cleanliness index and material stability score to generate a quality risk probability; The environmental cleanliness index is generated by a weighted comprehensive evaluation of multiple environmental parameters; the material stability score is the result of a weighted sum and normalization of material batch inspection data and supplier historical quality ratings.

4. The cloud-based blood sample collector production scheduling optimization system according to claim 1, characterized in that, The specific operation of the dynamic decision-making unit is as follows: When the overall production conflict index is less than or equal to the first decision threshold, the efficiency-first strategy is generated as the real-time optimal instruction. When the overall production conflict index is greater than the first decision threshold and less than or equal to the second decision threshold, a balance strategy is generated as the real-time optimal instruction. When the overall production conflict index exceeds the second decision threshold, a quality-first strategy is generated as the real-time optimal instruction.

5. The cloud-based blood sample collector production scheduling optimization system according to claim 1, characterized in that, The risk prediction unit uses a long short-term memory network model to learn a comprehensive state vector sequence that includes a historical comprehensive conflict index sequence, key physical parameters, and supply chain parameters, so as to output a comprehensive state vector evolution sequence for multiple future time steps, and extract the fluctuation of future key physical parameters from the comprehensive state vector evolution sequence.

6. The cloud-based blood sample collector production scheduling optimization system according to claim 5, characterized in that, The process of generating the feedforward pre-compensation command is as follows: Based on the fluctuations of key physical parameters in the future, and using a pre-set process response surface model, the compensation amount of the process parameters is calculated to generate feedforward pre-compensation instructions.

7. The cloud-based blood sample collector production scheduling optimization system according to claim 1, characterized in that, The fusion control unit is used to quantize the real-time optimal command into a dimensionless real-time adjustment factor.

8. The cloud-based blood sample collector production scheduling optimization system according to claim 7, characterized in that, The fusion control unit is also used to apply the compensation amount contained in the feedforward pre-compensation command and the real-time adjustment factor together to the preset reference process parameters to generate composite control commands.

9. The cloud-based blood sample collector production scheduling optimization system according to claim 1, characterized in that, The comprehensive production conflict index is generated through a nonlinear weighted model, and its calculation formula is as follows: ,in, Efficiency pressure index. For the probability of quality risk, and These are the basic weighting coefficients for efficiency and quality, respectively.

10. The cloud-based blood sample collector production scheduling optimization system according to claim 2, characterized in that, The formula for calculating the efficiency pressure index is as follows: ,in, Efficiency pressure index. The normalized equipment utilization rate To normalize the order urgency, Assign weights to efficiency.

11. The cloud-based blood sample collector production scheduling optimization system according to claim 3, characterized in that, The formula for calculating the probability of quality risk is: For the probability of quality risk, The environmental cleanliness index, To score material stability, Assign weights to risks.

12. The cloud-based blood sample collector production scheduling optimization system according to claim 8, characterized in that, The formula for generating the composite control command is as follows: It is a composite control command. As the baseline process parameters, This refers to the compensation amount included in the feedforward pre-compensation instruction.

Images