An intelligent monitoring and control system for a production line for preparing a pharmaceutical intermediate

Abstract

The application discloses an intelligent monitoring and regulating system for a medicine intermediate preparation production line, and belongs to the technical field of industrial automation control.The system comprises a data acquisition module, a risk index calculation module, a cause-and-effect decoupling module and a decision output module.The system is used for identifying a process problem or a raw material problem as a current dominant risk according to a comparison result of a cause-and-effect contribution factor and a preset contribution threshold, and generating a differentiated intervention instruction matched with the dominant risk.The system not only acquires real-time temperature and concentration data, but also converts the data into two risk indexes with clear physical meanings, and can accurately quantify physical risks caused by process problems such as local hot spots or fluid dead zones through construction of a fluid dynamics instability index.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology, specifically to a monitoring and control system for an intelligent pharmaceutical intermediate preparation production line. Background Technology

[0002] In modern pharmaceutical manufacturing, the preparation of high-value, chiral pharmaceutical intermediates is a crucial step. The production process, especially the multi-step exothermic crystallization stage, requires extremely stringent control of process conditions. This process is characterized by complex system states, nonlinearity, multivariable coupling, and high sensitivity to minute perturbations.

[0003] Currently, process analysis techniques are widely used on production lines. For example, in-situ Raman spectroscopy or Fourier transform infrared spectroscopy are used to monitor the concentration of chemical components in real time; distributed fiber optic sensors are used to monitor the temperature field inside the reactor; and computational fluid dynamics simulation is often used to simulate the flow field and temperature field under ideal conditions offline.

[0004] However, existing technologies have significant limitations:

[0005] When a system experiences a critical transition that deviates from its preset trajectory, such as a sudden abnormal crystallization rate or a sharp increase in byproducts, the underlying cause may be very complex. For example, it may be due to trace amounts of non-target configuration isomers in the raw materials introducing unexpected catalytic side reaction pathways; or it may be due to uneven stirring causing local hot spots or fluid dead zones in the reactor, which in turn triggers kinetic instability of the main reaction. Existing technologies lack an effective model that can perform real-time, quantitative causal decoupling in these coupled, competing failure pathways.

[0006] Because the cause cannot be accurately attributed, operators' interventions are often empirical and undifferentiated. When faced with an anomaly, should the stirring rate be adjusted immediately to improve mixing, or should a quality risk warning for the batch of raw materials be triggered immediately and the reaction be terminated? Incorrect decisions may not only lead to the scrapping of the entire batch of high-value products, but may even cause safety accidents such as thermal runaway.

[0007] 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

[0008] This invention aims to solve the technical challenge of real-time, quantitative causal decoupling of critical system transitions caused by the coupling effect of raw material quality and process conditions in the preparation of intelligent pharmaceutical intermediates, a problem that cannot be addressed in existing technologies. This invention constructs an indivisible intelligent analysis system that integrates multi-source data acquisition, quantitative index calculation, and differentiated decision output. This system identifies the decline in system dynamic resilience in real time and clearly determines the dominant factors leading to this change, providing a scientific basis for adopting optimal intervention strategies.

[0009] The technical solution of this invention is: a monitoring and control system for an intelligent pharmaceutical intermediate preparation production line, comprising:

[0010] Data acquisition module: used to collect real-time temperature data and real-time chemical component concentration data of the production process, and to construct a process status dataset based on the real-time temperature data and the real-time chemical component concentration data;

[0011] Risk index calculation module: used to perform the following steps:

[0012] Step 1: Calculate the fluid dynamics instability index based on the real-time temperature data in the process state dataset and the preset reference temperature data;

[0013] Step 2: Calculate the isomer catalytic risk index based on the real-time chemical component concentration data in the process state dataset;

[0014] Step 3: Weight and fuse the hydrodynamic instability index with the isomer catalytic risk index to generate the system dynamic toughness index;

[0015] Causal decoupling module: In response to events where the system dynamics toughness index is lower than a preset toughness threshold, it calculates the causal contribution factor for determining the dominant risk source by comparing the instantaneous change rate of the fluid dynamics instability index and the isomer catalytic risk index.

[0016] Decision output module: Based on the comparison results of causal contribution factors and preset contribution thresholds, it identifies process problems or raw material problems as the current dominant risks and generates differentiated intervention instructions that match the dominant risks.

[0017] In this embodiment, the risk index calculation module is specifically used for:

[0018] Based on the real-time temperature data collected by the data acquisition module and the preset reference temperature data, the weighted mean square error of the normalized difference between the real-time temperature and the reference temperature is calculated to generate the fluid dynamics instability index.

[0019] The reference temperature data was obtained by running offline computational fluid dynamics simulation software under ideal process conditions.

[0020] In this embodiment, the risk index calculation module is specifically used for:

[0021] Based on the real-time chemical component concentration data collected by the data acquisition module, the relative generation rate of characteristic byproducts and main products is calculated.

[0022] The relative generation rate is multiplied by a preset scaling factor to generate the isomer catalytic risk index, which is used to characterize the degree of threat posed by side reactions catalyzed by trace isomers to the main reaction.

[0023] The proportionality coefficient is a semi-empirical parameter calibrated by conducting a series of trace isomer doping experiments on a laboratory scale and performing linear regression fitting on the formation rates of the target product and by-products.

[0024] In this embodiment, the risk index calculation module is further used for:

[0025] The instantaneous comprehensive failure rate is obtained by multiplying the hydrodynamic instability index and the isomer catalytic risk index by their respective preset weighting coefficients and then summing them.

[0026] Based on the instantaneous comprehensive failure rate, the system dynamic resilience index is constructed through an exponential decay function to comprehensively evaluate the system's ability to resist risks and maintain stable operation.

[0027] The weighting coefficients are based on statistical analysis of historical data of a specific chemical process or are preset in combination with the knowledge of experts in the field, and are used to reflect the relative sensitivity of the process to the two types of risks.

[0028] In this embodiment, the causal decoupling module is specifically used for:

[0029] The instantaneous rate of change of the fluid dynamics instability index and the isomer catalytic risk index is calculated using their historical time series.

[0030] Multiply each instantaneous rate of change by its corresponding preset weighting coefficient to obtain the weighted rate of change;

[0031] The weight of the weighted rate of change of the fluid dynamic instability index in the sum of the absolute values ​​of the two weighted rates of change is calculated to determine the causal contribution factor.

[0032] In this embodiment, the decision output module is specifically used for:

[0033] When the causal contribution factor is higher than the preset first contribution threshold, the dominant cause of the current system instability is determined to be hydrodynamic non-uniformity, and the differentiated intervention instruction containing the instruction to increase the stirring rate or check the cooling system is generated.

[0034] In this embodiment, the decision output module is further specifically used for:

[0035] When the causal contribution factor is lower than the preset second contribution threshold, the dominant cause of the current system instability is determined to be the side reaction catalyzed by trace isomers, and the differentiated intervention instruction containing the triggering of a raw material batch quality alarm or the preparation of a safe shutdown instruction is generated.

[0036] In this embodiment, the decision output module is further specifically used for:

[0037] When the causal contribution factor is not lower than the second contribution threshold and not higher than the first contribution threshold, the system is determined to face a coupling problem and generates the differentiated intervention instruction that suggests comprehensive intervention measures.

[0038] In this embodiment, the data acquisition module is specifically used for:

[0039] The real-time temperature data is acquired by a distributed fiber Bragg grating sensor array deployed inside the reactor.

[0040] The real-time chemical component concentration data were calculated from the real-time spectrum using in-situ Raman spectroscopy combined with a partial least squares chemometrics model.

[0041] This invention provides an improved monitoring and control system for an intelligent pharmaceutical intermediate preparation production line, which has the following improvements and advantages compared with the prior art:

[0042] This system not only collects real-time temperature and concentration data, but also transforms this data into two risk indices with clear physical meanings. By constructing a fluid dynamics instability index, it can accurately quantify the physical risks caused by process problems such as local hot spots or fluid dead zones. By constructing an isomer catalysis risk index, it can quantify in real time the chemical risks of side reactions triggered by ppm-level impurities in raw materials. This dual-path quantitative assessment transforms the understanding of the production process from vague state anomalies to clear risk attribution.

[0043] The core advancement of this invention lies in its causal decoupling module. When the overall health of the system, i.e., the system dynamic resilience index, declines, the system does not issue a general alarm. Instead, it calculates causal contribution factors to diagnose the dominant risk source leading to system instability in real time and quantitatively. This achieves a leap from blind intervention to precise control. Based on clear root cause diagnosis, the decision output module can generate differentiated intervention instructions. If the diagnosis is a fluid dynamics problem, the instruction will focus on increasing the stirring rate or checking the cooling system; if the diagnosis is a trace isomer catalysis problem, the instruction will shift to triggering a raw material batch alarm or preparing for a safe shutdown. This precise matching control strategy avoids the ineffective operation of traditional control, significantly improves the success rate of intervention, and effectively reduces material loss and batch failure risks.

[0044] This system represents a leap from passive response to proactive early warning: by analyzing the instantaneous rate of change of the risk index to decouple the source of risk, it is far more sensitive than waiting for the risk index to exceed a fixed threshold. This dynamic trend-based judgment mechanism enables the system to identify the dominant direction of the problem in the early stages of risk accumulation, buying valuable response time for operators and transforming traditional passive emergency response into proactive risk intervention and critical transition early warning. Attached Figure Description

[0045] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0046] Figure 1 This is a flowchart of a monitoring and control system for an intelligent pharmaceutical intermediate preparation production line according to the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. Example

[0048] Please see Figure 1 This invention provides a technical solution for a monitoring and control system for an intelligent pharmaceutical intermediate preparation production line: Specific steps include:

[0049] Data acquisition module: Used to collect real-time temperature data and real-time chemical component concentration data of the production process, and to construct a process state dataset based on the real-time temperature data and real-time chemical component concentration data; The data acquisition module acquires temperature data in real time through a distributed fiber Bragg grating sensor array deployed in the reactor, and uses in-situ Raman spectroscopy combined with a partial least squares chemometrics model to calculate chemical component concentration data from the real-time spectrum, and constructs a process state dataset containing temperature field distribution and concentration change trajectory.

[0050] The risk index calculation module is used to perform the following steps:

[0051] Step 1: Calculate the fluid dynamics instability index based on the real-time temperature data in the process state dataset and the preset reference temperature data;

[0052] Step 2: Calculate the isomer catalytic risk index based on the real-time chemical component concentration data in the process state dataset;

[0053] Step 3: Weight and fuse the hydrodynamic instability index with the isomer catalytic risk index to generate the system dynamic toughness index;

[0054] The risk index calculation module generates a fluid dynamics instability index by calculating the weighted mean square error of the normalized difference between the real-time temperature and the reference temperature based on the real-time temperature data and the preset reference temperature data in the process state dataset. This index can accurately quantify the degree of deviation between the fluid dynamics state in the reactor and the ideal state. When local hot spots or fluid dead zones appear in the reactor, the fluid dynamics instability index increases significantly, indicating an increase in the risk of process problems.

[0055] Based on real-time chemical component concentration data in the process state dataset, an isomer catalytic risk index is generated by calculating the relative formation rate of characteristic byproducts and main products and multiplying it by a preset proportional coefficient. This index can quantify in real time the threat level of side reactions catalyzed by trace isomers to the main reaction. When there are ppm-level non-target configuration isomers in the feedstock, the isomer catalytic risk index rises sharply, indicating an aggravation of feedstock problem risks. The risk index calculation module weighted and fused the hydrodynamic instability index and the isomer catalytic risk index, and constructed a system kinetic toughness index through an exponential decay function. This index comprehensively evaluates the system's ability to resist risks and maintain stable operation. The closer the value is to 1, the better the system toughness; the lower the value, the closer the system is to the critical transition point.

[0056] Causal decoupling module: In response to events where the system dynamics toughness index is lower than a preset toughness threshold, it calculates the causal contribution factor for determining the dominant risk source by comparing the instantaneous change rate of the fluid dynamics instability index and the isomer catalytic risk index.

[0057] This causal contribution factor enables real-time, quantitative causal decoupling in coupled, competitive failure pathways. When the causal contribution factor approaches 1, it indicates that fluid dynamic inhomogeneity is the dominant causal factor. When the causal contribution factor approaches 0, it indicates that the side reaction catalyzed by trace isomers is the dominant causal factor. Based on the comparison between the causal contribution factor and the preset contribution threshold, the decision output module identifies process problems or raw material problems as the current dominant risks and generates differentiated intervention instructions that match the dominant risks. This achieves an upgrade from blind intervention to precise control, significantly improving production success rate and reducing material loss.

[0058] Decision output module: Based on the comparison results of causal contribution factors and preset contribution thresholds, it identifies process problems or raw material problems as the current dominant risks and generates differentiated intervention instructions that match the dominant risks.

[0059] In this embodiment, the risk index calculation module is specifically used for:

[0060] Based on the real-time temperature data collected by the data acquisition module and the preset reference temperature data, the weighted mean square error of the normalized difference between the real-time temperature and the reference temperature is calculated to generate the fluid dynamics instability index.

[0061] The reference temperature data was obtained by running offline computational fluid dynamics simulation software under ideal process conditions.

[0062] The following mathematical model is used to calculate the hydrodynamic instability index:

[0063]

[0064] Parameter explanation:

[0065] : The hydrodynamic instability index at time t; :time. Sensor spatial point index; Total number of temperature measurement points; : Always Real-time temperature collected at the point; : The offline CFD simulation reference temperature corresponding to the point; : Point weighting factors.

[0066] The reference temperature data was obtained through offline computational fluid dynamics simulation software under ideal process conditions. This reference temperature data represents the three-dimensional temperature field distribution under ideal steady-state operating conditions within the reactor. The fluid dynamics instability index can characterize the deviation of the fluid dynamics state within the reactor from the ideal state in real time. When uneven stirring leads to local hot spots or fluid dead zones within the reactor, the difference between the real-time temperature and the reference temperature increases significantly, and the fluid dynamics instability index rises accordingly, providing a quantitative basis for identifying process problems. This calculation method uses... The calculations implemented normalization, eliminating the influence of temperature dimensions and ensuring that the exponent is dimensionless; the weighted processing highlighted the importance of key monitoring points, achieving accurate quantification of fluid dynamic non-uniformity.

[0067] In this embodiment, the risk index calculation module is specifically used for:

[0068] Based on the real-time chemical component concentration data collected by the data acquisition module, the relative generation rate of characteristic byproducts and main products is calculated.

[0069] The relative generation rate is multiplied by a preset scaling factor to generate the isomer catalytic risk index, which is used to characterize the degree of threat posed by side reactions catalyzed by trace isomers to the main reaction.

[0070] The proportionality coefficient is a semi-empirical parameter calibrated by conducting a series of trace isomer doping experiments on a laboratory scale and performing linear regression fitting on the formation rates of the target product and by-products.

[0071] The isomer catalytic risk index is calculated using the following mathematical model:

[0072]

[0073] Specifically, : The risk index of isomer catalysis at any given time; :Concentration of main product; : Concentration of characteristic byproducts generated by isomer catalysis; The instantaneous rate of change of the concentration of the corresponding substance, i.e., the generation rate; : Proportional coefficient; : To prevent small non-negative constants with a denominator of zero.

[0074] proportionality coefficient The isomer catalysis risk index, a semi-empirical parameter calibrated by linear regression fitting of the formation rates of target products and byproducts through laboratory-scale trace isomer doping experiments, can quantify in real time the threat level of side reactions catalyzed by trace isomers to the main reaction. Its construction logic originates from the concept of selectivity in chemical reaction kinetics, defining the risk level by comparing the relative formation rates of target byproducts and main products. When non-target isomers at the ppm level are present in the feedstock, these isomers introduce unexpected catalytic side reaction pathways, leading to a significant increase in the formation rate of characteristic byproducts relative to the formation rate of the main product. The isomer catalysis risk index increases accordingly, providing a quantitative basis for identifying feedstock problems. This calculation method eliminates the influence of reaction scale through the comparison of relative formation rates, and the accuracy of risk assessment is ensured by calibrating the proportionality coefficient.

[0075] In this embodiment, the risk index calculation module is further used for:

[0076] The instantaneous comprehensive failure rate is obtained by multiplying the hydrodynamic instability index and the isomer catalytic risk index by their respective preset weighting coefficients and then summing them.

[0077] Based on the instantaneous comprehensive failure rate, the system dynamic resilience index is constructed through an exponential decay function to comprehensively evaluate the system's ability to resist risks and maintain stable operation.

[0078] The weighting coefficients are based on statistical analysis of historical data of a specific chemical process or are preset in combination with the knowledge of experts in the field, and are used to reflect the relative sensitivity of the process to the two types of risks.

[0079] The following mathematical model is used to calculate the system dynamics resilience index:

[0080]

[0081] Parameter explanation: : The system dynamics resilience index at time t is in the range of (0,1]; Weighting coefficients for fluid dynamics risk; Weighting coefficients for isomer risk; Fluid dynamics instability index; Isomer catalysis risk index.

[0082] The weighting coefficients are based on statistical analysis of historical data for specific chemical processes or pre-set using domain expert knowledge to reflect the relative sensitivity of the process to two types of risks. The system kinetic resilience index comprehensively assesses the system's ability to resist risks and maintain stable operation. Its mathematical form borrows from failure rate models in reliability engineering, treating the weighted sum of the two risk indices as the system's instantaneous comprehensive failure rate. The system kinetic resilience index ranges from (0,1), with values ​​closer to 1 indicating better system resilience and lower values ​​indicating the system is closer to the critical transition point. When the hydrodynamic instability index or isomer catalysis risk index increases, the instantaneous comprehensive failure rate increases, and the system kinetic resilience index decreases accordingly, providing a quantitative indicator for early warning of system critical transitions. This construction method achieves a nonlinear mapping from risk indices to resilience indices through an exponential decay function, and the weighting coefficients reflect the differences in sensitivity of different processes to different risk types.

[0083] In this embodiment, the causal decoupling module is specifically used for:

[0084] The instantaneous rate of change of the fluid dynamics instability index and the isomer catalytic risk index is calculated using their historical time series.

[0085] Multiply each instantaneous rate of change by its corresponding preset weighting coefficient to obtain the weighted rate of change;

[0086] The weight of the weighted rate of change of the fluid dynamic instability index in the sum of the absolute values ​​of the two weighted rates of change is calculated to determine the causal contribution factor.

[0087] The causal contribution factor is calculated using the following mathematical model:

[0088]

[0089] Parameter explanation: : Causal contribution factors at any given moment; The instantaneous rate of change of the fluid dynamics instability index; Instantaneous rate of change of the isomer catalytic risk index; , The weighting coefficients defined above.

[0090] The derivation of the causal contribution factor is based on the total differential of the system dynamics resilience index over time. After decomposition using the chain rule, the contribution ratio of each risk term's rate of change to the overall rate of change is calculated. The causal contribution factor's value range is defined as follows: when the causal contribution factor approaches 1, it indicates that the dominant causal factor for the current system instability is fluid dynamic inhomogeneity; when the causal contribution factor approaches 0, it indicates that the dominant causal factor is the side reaction catalyzed by trace isomers. The causal decoupling module achieves real-time quantitative decoupling of coupled risks by comparing the instantaneous rates of change of the two risk indices, solving the technical challenge of real-time quantitatively distinguishing whether process problems or raw material problems play a dominant role during system anomalies. This calculation method ensures the comparability of the rates of change of different risk types through the introduction of weighting coefficients, and achieves quantitative attribution of the dominant causal factor through weighting calculation.

[0091] In this embodiment, the decision output module is specifically used for:

[0092] When the causal contribution factor exceeds a preset first contribution threshold, the dominant causal factor for the current system instability is determined to be fluid dynamic inhomogeneity, and a differentiated intervention instruction is generated, including instructions to increase the stirring rate or check the cooling system. Upon receiving the value of the causal contribution factor, the decision output module immediately compares it with the first contribution threshold. When the causal contribution factor exceeds the first contribution threshold, the system determines that fluid dynamic inhomogeneity is the dominant risk source, indicating the presence of process problems such as local hot spots, fluid dead zones, or uneven mixing within the reactor. The differentiated intervention instruction provides precise control measures targeting the root cause of fluid dynamic inhomogeneity, including specific operational guidance such as immediately increasing the stirring rate to improve mixing, checking the cooling system's operating status to eliminate local hot spots, and adjusting the feed rate to optimize reaction kinetic balance.

[0093] In this embodiment, the decision output module is further specifically used for:

[0094] When the causal contribution factor is lower than the preset second contribution threshold, the dominant cause of the current system instability is determined to be the side reaction catalyzed by trace isomers, and a differentiated intervention instruction is generated, which includes triggering a batch quality alarm or preparing a safe shutdown instruction. After receiving the value of the causal contribution factor, the decision output module immediately compares it with the second contribution threshold. When the causal contribution factor is lower than the second contribution threshold, the system determines that the side reaction catalyzed by trace isomers is the dominant source of risk, indicating that there are excessive non-target configuration isomers in the current batch of raw materials. The differentiated intervention instruction provides emergency risk control measures for the side reaction catalyzed by trace isomers, including immediately triggering a quality alarm for the batch of raw materials to trace the source of the raw materials, suspending feeding to prevent the problem from escalating, and preparing a safe shutdown procedure to avoid thermal runaway and other key operational guidance.

[0095] In this embodiment, the decision output module is further specifically used for:

[0096] When the causal contribution factor is not lower than the second contribution threshold and not higher than the first contribution threshold, the system is determined to face a coupling problem, and a differentiated intervention instruction suggesting comprehensive intervention measures is generated. The decision output module identifies a complex situation where fluid dynamic inhomogeneity and the side reaction catalyzed by trace isomers simultaneously contribute significantly to system instability by accurately judging the numerical range of the causal contribution factor. At this time, a single intervention measure cannot effectively solve the multiple risks faced by the system. The differentiated intervention instruction provides a systematic and comprehensive control scheme for the complexity of the coupling problem, including adjusting the stirring rate and cooling system parameters to improve the fluid dynamic state, initiating a deep detection program for raw material quality to confirm the isomer content, and implementing phased process parameter optimization to balance the interaction between the two types of risks, etc., providing multi-dimensional operational guidance.

[0097] In this embodiment, the data acquisition module is specifically used for:

[0098] The real-time temperature data is acquired by a distributed fiber Bragg grating sensor array deployed inside the reactor.

[0099] The real-time chemical component concentration data is calculated from the real-time spectrum using in-situ Raman spectroscopy combined with a partial least squares chemometrics model. A distributed fiber Bragg grating sensor array is deployed at multiple points, with temperature monitoring points set at key locations within the reactor, including near the feed inlet, around the stir bar, at the reactor wall edge, and at the bottom—locations prone to temperature gradients—to achieve comprehensive monitoring of the three-dimensional temperature field within the reactor. The in-situ Raman spectroscopy system obtains chemical bond vibration information through the Raman scattering effect of sample molecules excited by laser. The partial least squares chemometrics model establishes a quantitative relationship between spectral characteristics and chemical component concentrations, enabling accurate conversion from spectral data to concentration data.

[0100] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A monitoring and regulating system for an intelligent pharmaceutical intermediate preparation production line, characterized in that, The method comprises the following steps: a data acquisition module is configured to collect real-time temperature data and real-time chemical component concentration data of a production process, and construct a process state data set based on the real-time temperature data and the real-time chemical component concentration data; a risk index calculation module is configured to perform the following steps: Step 1: calculating a fluid dynamics instability index based on real-time temperature data in the process state data set and preset reference temperature data; Step 2: calculating an isomer catalysis risk index based on real-time chemical component concentration data in the process state data set; Step 3: weighting and fusing the fluid dynamics instability index and the isomer catalysis risk index to generate a system dynamics resilience index; a cause-and-effect decoupling module is configured to, in response to an event that the system dynamics resilience index is lower than a preset resilience threshold, calculate a cause-and-effect contribution factor for determining a dominant risk source by comparing instantaneous change rates of the fluid dynamics instability index and the isomer catalysis risk index; a decision output module is configured to identify a process problem or a raw material problem as a current dominant risk according to a comparison result of the cause-and-effect contribution factor and a preset contribution threshold, and generate a differentiated intervention instruction matched with the dominant risk; the risk index calculation module is specifically configured to: calculate a weighted mean square error of a normalized difference between the real-time temperature data collected by the data acquisition module and the preset reference temperature data to generate the fluid dynamics instability index; wherein the reference temperature data is obtained by running offline computational fluid dynamics simulation software to simulate and calculate under ideal process conditions; the risk index calculation module is specifically configured to: calculate a relative generation rate of a characteristic byproduct and a main product based on the real-time chemical component concentration data collected by the data acquisition module; multiply the relative generation rate by a preset proportionality coefficient to generate the isomer catalysis risk index, which is used to represent a threat degree of a by-reaction catalyzed by a trace isomer to a main reaction; wherein the proportionality coefficient is a semi-empirical parameter calibrated by performing a series of trace isomer doping experiments at a laboratory scale and performing linear regression fitting on generation rates of target products and byproducts. 2.The monitoring and regulating system for intelligent pharmaceutical intermediate preparation and production line according to claim 1, characterized in that, the risk index calculation module is further configured to: sum the fluid dynamics instability index and the isomer catalysis risk index after multiplying them by their respective preset weight coefficients to obtain an instantaneous comprehensive failure rate; construct the system dynamics resilience index by an exponential decay function based on the instantaneous comprehensive failure rate, which is used to comprehensively evaluate the ability of the system to resist risks and maintain stable operation; wherein the weight coefficients are preset based on statistical analysis of historical data of a specific chemical process or combined with domain expert knowledge, and are used to reflect the relative sensitivity of the process to the two types of risks. 3.The monitoring and regulating system for intelligent pharmaceutical intermediate preparation and production line according to claim 2, characterized in that, the cause-and-effect decoupling module is specifically configured to: calculate instantaneous change rates of the fluid dynamics instability index and the isomer catalysis risk index using their historical time series; multiply each instantaneous change rate by its corresponding preset weight coefficient to obtain a weighted change rate; The proportion of the weighted change rate of the fluid dynamics instability index in the sum of absolute values of two weighted change rates is calculated to obtain the causal contribution factor. 4.The monitoring and regulating system for intelligent pharmaceutical intermediate preparation and production line according to claim 3, characterized in that, The decision output module is specifically configured to: When the causal contribution factor is higher than a preset first contribution threshold, it is determined that the dominant cause of the current system instability is fluid dynamics heterogeneity, and the differentiated intervention instruction containing an instruction to increase the stirring rate or check the cooling system is generated. 5.The monitoring and regulating system for intelligent pharmaceutical intermediate preparation and production line according to claim 4, characterized in that, The decision output module is further specifically configured to: When the causal contribution factor is lower than a preset second contribution threshold, it is determined that the dominant cause of the current system instability is a trace isomer catalyzed side reaction, and the differentiated intervention instruction containing an instruction to trigger a raw material batch quality alarm or prepare a safety shutdown instruction is generated. 6.The monitoring and regulating system for intelligent pharmaceutical intermediate preparation and production line according to claim 5, characterized in that, The decision output module is further specifically configured to: When the causal contribution factor is not lower than the second contribution threshold and not higher than the first contribution threshold, it is determined that the system is facing a coupling problem, and the differentiated intervention instruction suggesting to take comprehensive intervention measures is generated. 7.The intelligent monitoring and regulation system for the production line of pharmaceutical intermediates according to claim 1, characterized in that, The data acquisition module is specifically configured to: The real-time temperature data is acquired through a distributed fiber Bragg grating sensor array deployed in the reaction kettle; The real-time chemical component concentration data is calculated from the real-time spectrum through in-situ Raman spectrum combined with a partial least squares chemometrics model.

Images