Continuous flow chemical product carbon tracking method and system based on rtd dynamic compensation

Abstract

The application discloses a continuous fluid chemical product carbon tracking method and system based on RTD dynamic compensation, comprising: reactor RTD mathematical characterization modeling and generating a probability density sequence; the reactor is abstracted into a one-stage or multi-stage series stirred tank model, and a residence time distribution density function is used to describe the residence probability of fluid microgroups in the reactor; an instantaneous carbon load factor is constructed; reactor inlet flow, raw material implicit carbon intensity and multi-path energy consumption are collected in real time, and dispersed energy consumption and material data are integrated into a carbon flow signal changing with time; and output carbon accounting based on time domain convolution. The application applies mixing theory in chemical engineering to carbon accounting, realizes a leap from average carbon to instantaneous carbon, and can provide a precise carbon single with a time stamp for downstream customers.

Description

Technical Field

[0001] This invention relates to the fields of carbon emission accounting and chemical process control, specifically to an instantaneous carbon footprint tracking technology for continuous flow reaction systems (such as stirred tanks, wastewater treatment ponds, etc.). Background Technology

[0002] Carbon footprint accounting has become an important basis for assessing the environmental impact of products and formulating emission reduction strategies.

[0003] The following existing technologies were found through a search: The patent specification with publication number CN116468204A discloses a method, system, device, and storage medium for online monitoring of the carbon footprint of industrial products. The method includes: collecting energy flow data and material flow data during the product production process within a collection period; constructing a product carbon footprint calculation model based on the life cycle assessment methodology to calculate the product carbon footprint data; setting a carbon footprint alarm threshold, determining whether the product carbon footprint data triggers a carbon footprint alarm mechanism, generating an analysis report, and generating a product carbon footprint management instruction; and performing carbon footprint monitoring and management of the product production process based on the product carbon footprint management instruction.

[0004] The patent specification with publication number CN118798914A discloses a method for optimizing carbon emissions calculation in a chemical plant. This method includes: establishing a mathematical model simulating the process flow of the target chemical plant based on its actual operating parameters; performing process calculations on the target chemical plant based on the mathematical model, classifying and calculating material and energy flows according to the calculation results to obtain a set of carbon emission sources corresponding to the target chemical plant; the set of carbon emission sources includes carbon emission sources caused by material flow and / or carbon emission sources caused by energy flow; setting objective functions, constraints, and decision variables, establishing a multi-objective optimization model for chemical plant production based on parameter tuning, solving for the optimized production scheme, and thus calculating the final carbon emissions.

[0005] In continuous flow chemical production, back-mixing occurs within the reactor, resulting in a high degree of spatial mixing between new and old materials. Existing accounting methods (such as periodic averaging) cannot distinguish the differentiated contributions of energy sources entering at different times (such as green electricity versus thermal power) to the carbon footprint of the final product, leading to low accounting accuracy and an inability to support the traceability requirements of high-value green products. In other words, the mixing of new and old materials in a continuous flow reactor makes it difficult to accurately determine which energy source entered at a specific point in time was consumed by the product produced at that particular time (e.g., different peak and off-peak electricity prices lead to different carbon emission factors). Summary of the Invention

[0006] Traditional methods allocate carbon emissions on a monthly / annual average basis, which cannot distinguish the carbon differences between different batches of products. This invention introduces Residence Time Distribution (RTD) and provides a continuous flow chemical product carbon tracking method and system based on RTD dynamic compensation. It can solve the carbon footprint ambiguity problem caused by backmixing and calculate the true carbon footprint of the product produced at the current moment.

[0007] The specific technical solution is as follows: In a first aspect, the present invention provides a continuous flow chemical product carbon tracking method based on RTD dynamic compensation, comprising: Reactor RTD mathematical representation modeling and probability density sequence generation: The reactor is abstracted into a single-stage or multi-stage series stirred tank model (N-CSTR), and the residence time distribution density function is used to describe the residence probability of fluid particles in the reactor; Constructing an instantaneous carbon loading factor: Real-time acquisition of reactor inlet flow rate, raw material embodied carbon intensity, and multi-channel energy consumption, integrating dispersed energy consumption and material data into a carbon flow signal that varies over time; Carbon production calculation based on temporal convolution.

[0008] In some embodiments, the residence time distribution density function is expressed as follows:

[0009] This represents the residence time of fluid particles in the reactor (independent variable). This indicates the average length of stay. , Indicates volume. Indicates flow rate; For model series, This represents complete backmixing (ideal stirred tank). This represents a completely non-remixing (flat flow); Represents probability density; This indicates that the product has remained in the output. Material proportions over time.

[0010] The reactor can be modeled physically (e.g., using CFD simulation) or through tracer experiments (e.g., injecting tracers) to acquire reactor outlet concentration curves. Based on the reactor outlet concentration and the N-CSTR model formula, the optimal value can be fitted using the least squares method. and value.

[0011] In some embodiments, the instantaneous carbon load factor is constructed using the following formula:

[0012] express Total carbon loading flux entering the reactor at any given time ( ), namely, instantaneous carbon loading factor; express Instantaneous feed volume flow rate of reactor at any time ( ); express The carbon intensity of raw materials at any time ( ); Indicates the density of the material; express Time of the first Instantaneous electrical power of equipment (such as stirring motor, heater, etc.) ); Indicates the first The effective energy efficiency ratio of the equipment (excluding reactive power); express Real-time dynamic grid emission factor.

[0013] In some embodiments, a length of [length] is established. A ring-shaped buffer is used to store the instantaneous carbon load factor in real time; Align the probability density sequence with the instantaneous carbon load factor weights stored in the circular buffer; The carbon accounting based on temporal convolution includes: reading several instantaneous carbon load factors from the current time to the previous time in the circular buffer, multiplying the instantaneous carbon load factors corresponding to the same time with the probability density point by point, then summing all the product results and dividing by the production rate at the current time. The circular buffer slides through the update process, which involves kicking out the oldest data and pushing in the latest instantaneous carbon load factor to recalculate the produced carbon based on temporal convolution.

[0014] In some embodiments, the following formula is used for calculating the produced carbon based on temporal convolution:

[0015] Indicates the current moment; This represents the carbon footprint per unit weight of product produced at the current moment (in units). (That is, how much carbon dioxide is emitted per kilogram of product). This represents the instantaneous mass flow rate at the output end at the current moment. ); This indicates the backtracking cutoff point of the convolution; Let be the independent variable of the convolution integral, representing the time step of backtracking, for example, when At that time, the above formula is used to process materials that entered the reactor 10 seconds ago; Indicates the time starting from the current moment Total carbon loading flux entering the reactor at a time prior to time ( ); The residence time distribution density function reflects the time from the current moment. The probability that material that entered at a time before the specified time will be discharged at the current time.

[0016] In some embodiments, the continuous flow chemical product carbon tracking method based on RTD dynamic compensation further includes feedback correction, that is, using correction gain to compensate for the final carbon footprint, specifically using the following formula:

[0017]

[0018] This represents the final carbon footprint per unit weight of product produced at the current moment after correction. The feedback gain coefficient, similar to the proportional gain in PID (Proportional-Integral-Derivative) control, determines the strength of the correction. This indicates the measured concentration at the reactor outlet at the current moment, obtained through sensor measurement. This indicates the model-predicted concentration at the reactor outlet at the current moment; The carbon-matter conversion factor is the conversion coefficient that transforms material concentration deviation into carbon footprint deviation. Indicates the time starting from the current moment The concentration of the material entering the reactor at a time prior to the specified time.

[0019] In some embodiments, the continuous flow chemical product carbon tracking method based on RTD dynamic compensation is subject to the following constraints: Gain saturation limit: The absolute value of the term and When the ratio exceeds the set threshold (e.g., 20%), an error is reported, which can be determined as a sensor malfunction or a major change in the process. Dead zone protection: when When the absolute value of an item is less than the sensor's measurement accuracy limit (such as 1% of full scale), no compensation is performed to prevent high-frequency numerical oscillations. Convergence criterion: If the feedback correction direction is opposite for multiple consecutive sampling periods (e.g., 10 times), oscillation occurs, and a forced lock is triggered. It also suggests that manual intervention be performed.

[0020] In some embodiments, the continuous flow chemical product carbon tracking method based on RTD dynamic compensation further includes flow field offset diagnosis (which can be triggered periodically) for optimizing the model level; The following variance criterion formula is used for flow field migration diagnosis:

[0021] The variance value measures the degree of deviation between the actual residence time distribution and the theoretical model. Indicates the sampling time; Indicates the sampling period; This indicates the average length of stay. , Indicates volume. Indicates flow rate; The measured concentration sequence at the reactor outlet can be tracer data obtained by online sensors (such as conductivity meters). For model series; The model theory has dimensionless variance; When the value exceeds a preset threshold (e.g., 0.1), adjust the model level. ,make Reduce to no more than a preset threshold. If the variance value An increase indicates enhanced back-mixing, which automatically reduces the number of model levels. value.

[0022] Secondly, the present invention provides a continuous flow chemical product carbon tracking system based on RTD dynamic compensation, comprising: RTD modeling unit is used for RTD mathematical characterization modeling of reactor and generation of probability density sequence: the reactor is abstracted into a single-stage or multi-stage series stirred tank model, and the residence time distribution density function is used to describe the residence probability of fluid particles in the reactor. The carbon load construction unit is used to construct the instantaneous carbon load factor: real-time acquisition of reactor inlet flow, raw material implicit carbon intensity and multiple energy consumption, integrating the scattered energy consumption and material data into a carbon flow signal that varies over time. A convolutional kernel engine for producing carbon based on temporal convolution.

[0023] In some embodiments, the continuous flow chemical product carbon tracking system based on RTD dynamic compensation further includes: The feedback correction unit is used to compensate for the final carbon footprint and diagnose flow field deviations to optimize the model series.

[0024] Thirdly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the computer program to implement the steps of the continuous flow chemical product carbon tracking method based on RTD dynamic compensation as described in the first aspect.

[0025] Compared with the prior art, the beneficial effects of this invention are as follows: This invention applies the mixing theory from chemical engineering to carbon accounting, achieving a leap from average carbon to instantaneous carbon, and can provide downstream customers with accurate carbon statements with timestamps. Attached Figure Description

[0026] Figure 1 This is a diagram illustrating the architecture of a continuous flow chemical product carbon tracking system based on RTD dynamic compensation according to the present invention.

[0027] Figure 2 This is a diagram illustrating the effect of real-time carbon footprint tracking and feedback correction. Detailed Implementation

[0028] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0029] See Figure 1 A continuous flow chemical product carbon tracking system based on RTD dynamic compensation includes: The RTD modeling unit (module 1) is used for RTD mathematical characterization modeling of reactors and generation of probability density sequences: it establishes an RTD model of non-ideal flow in the reactor, abstracts the reactor into a single-stage or multi-stage series stirred tank model, uses the residence time distribution density function to describe the residence probability of fluid particles in the reactor, and determines the residence law of materials in the stirred tank (CSRT). The carbon load construction unit (module 2) is used to construct the instantaneous carbon load factor: real-time acquisition of reactor inlet flow, raw material implicit carbon intensity and multi-channel energy consumption, integrating the dispersed energy consumption and material data into a carbon flow signal that changes over time. Convolutional kernel calculation engine (Module 3) is used for carbon production calculation based on temporal convolution; The feedback correction unit (module four) is used to compensate for the final carbon footprint and diagnose flow field shifts to optimize the model series by utilizing correction gains.

[0030] A continuous flow chemical product carbon tracking method based on RTD dynamic compensation, employing the aforementioned continuous flow chemical product carbon tracking system based on RTD dynamic compensation, includes: Reactor RTD mathematical representation modeling and probability density sequence generation: Establish the RTD model of non-ideal flow in the reactor, abstract the reactor into a single-stage or multi-stage series stirred tank model (N-CSTR), use the residence time distribution density function to describe the residence probability of fluid particles in the reactor, and determine the residence law of materials in the stirred tank (CSRT).

[0031] The residence time distribution density function is expressed as follows (which can be represented by the N-CSTR formula):

[0032] This represents the residence time of fluid particles in the reactor (independent variable). This indicates the average length of stay. , Indicates volume. Indicates flow rate; For model series, This represents complete backmixing (ideal stirred tank). This represents a completely non-remixing (flat flow); Represents probability density; This indicates that the product has remained in the output. Material proportions over time.

[0033] The reactor can be modeled physically (e.g., using CFD simulation) or through tracer experiments (e.g., injecting tracers) to acquire reactor outlet concentration curves. Based on the reactor outlet concentration and the N-CSTR model formula, the optimal value can be fitted using the least squares method. and value.

[0034] Constructing an instantaneous carbon loading factor: Real-time acquisition of reactor inlet flow rate, raw material embodied carbon intensity, and multi-channel energy consumption, integrating dispersed energy consumption and material data into a carbon flow signal that varies over time.

[0035] The instantaneous carbon load factor is constructed using the following formula:

[0036] express Total carbon loading flux entering the reactor at any given time ( ), namely, instantaneous carbon loading factor; express Instantaneous feed volume flow rate of reactor at any time ( ); express The carbon intensity of raw materials at any time ( ); Indicates the density of the material; express Time of the first Instantaneous electrical power of equipment (such as stirring motor, heater, etc.) ); Indicates the first The effective energy efficiency ratio of the equipment (excluding reactive power); express Real-time dynamic grid emission factor (this value is updated every minute and reflects changes in the proportion of green electricity).

[0037] Carbon production calculation based on temporal convolution.

[0038] Create a length of A ring-shaped buffer is used to store the instantaneous carbon load factor in real time; Align the probability density sequence with the instantaneous carbon load factor weights stored in the circular buffer; the probability density sequence can be represented as ; The carbon accounting based on temporal convolution includes: reading several instantaneous carbon load factors from the current time to the previous time in the circular buffer, multiplying the instantaneous carbon load factors corresponding to the same time with the probability density point by point, then summing all the product results and dividing by the production rate at the current time. The circular buffer slides through the update process, which involves kicking out the oldest data and pushing in the latest instantaneous carbon load factor to recalculate the produced carbon based on temporal convolution.

[0039] The following formula is used for calculating the produced carbon based on temporal convolution:

[0040] Indicates the current moment; This represents the carbon footprint per unit weight of product produced at the current moment (in units). (That is, how much carbon dioxide is emitted per kilogram of product). This represents the instantaneous mass flow rate at the output end at the current moment. ); The backtracking cutoff point of the convolution can typically be set to 0. ~ (That is, 3 to 5 times the average dwell time). After this time, the old materials have been almost completely emptied, and the probability is close to 0. There is no need to calculate anymore, thus saving computing power (loss stop design). Let be the independent variable of the convolution integral, representing the time step of backtracking, for example, when At that time, the above formula is used to process materials that entered the reactor 10 seconds ago; Indicates the time starting from the current moment Total carbon loading flux entering the reactor at a time prior to time ( ); The residence time distribution density function reflects the time from the current moment. The probability that material that entered at a time before the specified time will be discharged at the current time.

[0041] Feedback correction, which uses correction gain to compensate for the final carbon footprint, is specifically expressed by the following formula:

[0042]

[0043] This represents the final carbon footprint per unit weight of product produced at the current moment after correction. The feedback gain coefficient, similar to the proportional gain in PID (Proportional-Integral-Derivative) control, determines the strength of the correction. This indicates the measured concentration at the reactor outlet at the current moment, obtained through sensor measurement. This indicates the model-predicted concentration at the reactor outlet at the current moment; The carbon-matter conversion factor is the conversion coefficient that transforms material concentration deviation into carbon footprint deviation. Indicates the time starting from the current moment The concentration of the material entering the reactor at a time prior to the specified time.

[0044] The following conditions are applied: Gain saturation limit: The absolute value of the term and When the ratio exceeds the set threshold (e.g., 20%), an error is reported, which can be determined as a sensor malfunction or a major change in the process. Dead zone protection: when When the absolute value of an item is less than the sensor's measurement accuracy limit (such as 1% of full scale), no compensation is performed to prevent high-frequency numerical oscillations. Convergence criterion: If the feedback correction direction is opposite for multiple consecutive sampling periods (e.g., 10 times), oscillation occurs, and a forced lock is triggered. It also suggests that manual intervention be performed.

[0045] Flow field migration diagnostics (can be triggered periodically) are used to optimize model levels; The following variance criterion formula is used for flow field migration diagnosis:

[0046] The variance value measures the degree of deviation between the actual residence time distribution and the theoretical model. Indicates the sampling time; Indicates the sampling period; This indicates the average length of stay. , Indicates volume. Indicates flow rate; The measured concentration sequence at the reactor outlet can be tracer data obtained by online sensors (such as conductivity meters). For model series; The model theory has dimensionless variance; When the threshold value exceeds a preset threshold (e.g., 0.1), it indicates that a dead zone or short circuit has occurred inside the reactor, and the model level should be adjusted. ,make Reduce to no more than a preset threshold. For example, if the variance value... An increase indicates enhanced back-mixing, which automatically reduces the number of model levels. value.

[0047] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the computer program to implement the steps of the continuous flow chemical product carbon tracking method based on RTD dynamic compensation as described above.

[0048] Figure 2 The invention demonstrates the real-time carbon footprint tracking and feedback correction effect, and compares it with the traditional carbon footprint averaging method.

[0049] This document presents a specific application case of the above-mentioned continuous flow chemical product carbon tracking method and system based on RTD dynamic compensation, used in a dynamic carbon tracking system for a certain continuous flow acrylic acid polymerization process.

[0050] This case study uses the acrylic acid polymerization reaction system of a fine chemical enterprise with an annual output of 50,000 tons as a background to illustrate in detail the application logic, calculation process and beneficial effects of the method of the present invention in a real and complex production environment.

[0051] 1. Characterization of physical device parameters and initialization of RTD model: The acrylic acid polymerization process takes place in a controlled continuous flow stirred tank reactor (CSTR).

[0052] Physical parameter settings: Total effective volume of reactor =40m 3 Under standard operating conditions, the material feed flow rate =2m 3 / h, average density of material =1.05×10 3 kg / m 3 The system collects traffic and operating condition data every 5 minutes through the DCS interface and stores it in a real-time database to ensure that the calculation step size matches the physical response time.

[0053] RTD Model Fitting: Considering the arrangement of the agitators and the influence of internal components (such as heat exchange coils) within the reactor, the material flow pattern is between complete backmixing and plug flow. Experimental data were obtained by instantaneously injecting a pulse tracer (such as a 10% potassium chloride solution) at the inlet and monitoring the change in outlet conductivity online. The experimental curves were fitted to the N-CSTR formula using the least squares method to determine the model order. =4, correlation coefficient R 2 The value >0.98 verifies the high consistency between the model and the physical flow characteristics of the actual device.

[0054] Time scale definition: Calculated average stay time =20 hours. The system sets the sampling period to 5 minutes to generate a discretized probability density sequence, which serves as the core kernel function for subsequent convolution operations.

[0055] 2. Real-time carbon load monitoring and signal integration: The system connects to the enterprise energy management gateway (EMS) and DCS system (distributed control system) to collect production process data in real time. A typical sampling point is defined as 14:00 on May 10, 2024. Material side: Real-time flow =2m 3 / h. The implicit carbon emission factor of the raw material acrylic acid monomer was obtained from the supplier's carbon footprint report and set as the implicit carbon intensity of the raw material = 1.8 kgCO2 / kg.

[0056] Energy side: The total real-time power consumption (instantaneous power) of the reactor's main stirring motor and supporting refrigeration pump is 150kW.

[0057] Fluctuations in carbon intensity of electricity: This period coincides with the peak of local photovoltaic power generation, and the real-time emission factor (dynamic grid emission factor) of the power dispatch center has dropped to 0.45 kgCO2 / kWh (a 30% decrease compared to the nighttime average).

[0058] Calculation: Substitution The formula is used to calculate the total carbon loading flux entering the system at that moment: =(2×1.8×1050)+(150×0.45)=3847.5kgCO2 / h.

[0059] 3. Dynamic tracking and calculation: Temporal convolution and material balance: Due to the backmixing characteristics of continuous flow systems, the product discharged at 14:00 is actually a mixture of materials that entered the system over the past few dozen hours.

[0060] Convolution operation: The system retrieves data from the past 60 hours (approximately 3 times the average dwell time). Sequence. This involves matching the carbon load signals that entered at historical moments with the corresponding... The weights are accumulated to perform carbon accounting based on temporal convolution: the 60-hour historical window is discretized into 720 time steps (5 minutes per step). By discretizing and summing the carbon load signal input at each step with the corresponding weights, a numerical approximation of the continuous integral is achieved.

[0061] Technical implications: This processing method avoids the accounting distortion caused by traditional instantaneous methods when there are fluctuations in flow rate or sudden changes in energy intensity, and accurately restores the dynamic diffusion process of carbon signal in the reactor.

[0062] 4. Feedback compensation and abnormal operating condition constraints: To address the issue of sensor drift or model accuracy being affected by temperature and viscosity fluctuations, the system introduces a measured concentration feedback logic.

[0063] Deviation detection: At the outlet, the concentration of solids calculated by the online refractometer at the current moment (actual concentration at reactor outlet) is 0.85 mol / L, while the RTD predicted concentration (model predicted concentration at reactor outlet) is 0.82 mol / L.

[0064] Feedback correction: Set the feedback gain coefficient =0.5, carbon-matter conversion factor =1.2. According to Correction formula: = +0.5×(0.85-0.82)×1.2.

[0065] Safety constraints: Dead zone detection (dead zone protection): If the deviation value is 0.03, which is greater than the sensor's 1% measurement error limit, correction is triggered.

[0066] Gain saturation limit: The calculated correction term accounts for 4.5% of the original value, which does not exceed the alarm threshold of 20%, indicating that the system is operating stably and the correction is effective.

[0067] Convergence / Oscillation Detection: The system records the correction direction for 10 consecutive sampling periods. If the correction amount is detected to frequently jump between positive and negative ranges (e.g., 6 out of 10 times the direction is opposite), the system is determined to be experiencing numerical oscillation. In this case, the system is forcibly locked. It also triggers a manual intervention alarm to ensure that the calculation results are not distorted due to excessive control gain.

[0068] 5. Flow field migration diagnosis and Value dynamic optimization: During the production process, the flow field inside the reactor may shift due to increased material viscosity or slight adjustments in stirring speed.

[0069] Diagnostic logic: The system uses the variance criterion formula to determine the dimensionless variance of the outlet concentration distribution. Real-time monitoring is conducted.

[0070] Offset determination: initial settings =4, corresponding to a theoretical variance of 0.25. If the current measured variance shifts to 0.33, the system determines that the flow field is shifting towards strong backmixing.

[0071] Adaptive optimization: The system optimizes according to the relational formula. =1 / The model level was automatically adjusted from 4 to 3, and the residence time distribution density function was regenerated. sequence.

[0072] Results: The optimized model better reflects the current actual flow conditions, reducing the basic deviation between the predicted concentration and the measured value by 18%, thus improving the physical accuracy of carbon tracing from the source.

[0073] 6. Implementation Results and Management Value Analysis: By introducing the dynamic tracking method and system of this invention, the chemical company achieved the following beneficial effects: Product carbon label differentiation: Successfully identifying and labeling low-carbon batches of products produced between 2:00 PM and 4:00 PM (by utilizing the green radio off-peak hours) resulted in a 12% reduction in unit carbon footprint compared to the daily average. This enables companies to offer premium-priced low-carbon product certifications to high-end green supply chain customers.

[0074] Improved accounting accuracy: Compared with the traditional monthly material allocation method (total energy consumption / total output), this method reduces the accounting deviation from 15.6% to less than 2.4% during periods of drastic fluctuations in operating conditions (such as start-up, shutdown, or load adjustment).

[0075] Real-time emissions reduction decision-making: through With dynamic optimization of values, the system can not only accurately calculate the past but also predict the output carbon footprint trend in the next two hours, providing data support for operators to adjust the feed rate or stagger energy use, significantly enhancing the company's core competitiveness in the context of carbon neutrality.

[0076] Furthermore, it should be understood that after reading the above description of the present invention, those skilled in the art can make various alterations or modifications to the present invention, and these equivalent forms also fall within the scope defined by the appended claims.

Claims

1. A continuous flow chemical product carbon tracking method based on RTD dynamic compensation, characterized in that, include: Reactor RTD mathematical representation modeling and generation of probability density sequence: The reactor is abstracted into a single-stage or multi-stage series stirred tank model, and the residence time distribution density function is used to describe the residence probability of fluid particles in the reactor; Constructing an instantaneous carbon loading factor: Real-time acquisition of reactor inlet flow rate, raw material embodied carbon intensity, and multi-channel energy consumption, integrating dispersed energy consumption and material data into a carbon flow signal that varies over time; Carbon production calculation based on temporal convolution.

2. The continuous flow chemical product carbon tracking method based on RTD dynamic compensation according to claim 1, characterized in that, The residence time distribution density function is expressed as follows: ； This indicates the residence time of fluid particles within the reactor; This indicates the average length of stay. , Indicates volume. Indicates flow rate; For model series, This represents complete remixing. This means there is absolutely no mixing. Represents probability density; This indicates that the product has remained in the output. Material proportions over time.

3. The continuous flow chemical product carbon tracking method based on RTD dynamic compensation according to claim 1, characterized in that, The instantaneous carbon load factor is constructed using the following formula: ； express The total carbon loading flux entering the reactor at any given time, i.e., the instantaneous carbon loading factor; express Instantaneous feed volume flow rate of the reactor at any given moment; express The carbon intensity of raw materials at any given time; Indicates the density of the material; express Time of the first The instantaneous electrical power of the equipment; Indicates the first The effective energy efficiency ratio of the equipment; express Real-time dynamic power grid emission factor.

4. The continuous flow chemical product carbon tracking method based on RTD dynamic compensation according to claim 1, characterized in that, Create a length of A ring-shaped buffer is used to store the instantaneous carbon load factor in real time; Align the probability density sequence with the instantaneous carbon load factor weights stored in the circular buffer; The carbon accounting based on temporal convolution includes: reading several instantaneous carbon load factors from the current time to the previous time in the circular buffer, multiplying the instantaneous carbon load factors corresponding to the same time with the probability density point by point, then summing all the product results and dividing by the production rate at the current time. The circular buffer slides through the update process, which involves kicking out the oldest data and pushing in the latest instantaneous carbon load factor to recalculate the produced carbon based on temporal convolution.

5. The continuous flow chemical product carbon tracking method based on RTD dynamic compensation according to claim 1 or 4, characterized in that, The following formula is used for calculating the produced carbon based on temporal convolution: ； Indicates the current moment; This represents the carbon footprint per unit weight of product produced at the current moment; This represents the instantaneous mass flow rate at the output end at the current moment; This indicates the backtime cutoff point of the convolution; The independent variable of the convolution integral represents the time step of backtracking; Indicates the time starting from the current moment The total carbon loading flux entering the reactor at the time prior to the specified time. The residence time distribution density function reflects the time from the current moment. The probability that material that entered at a time before the specified time will be discharged at the current time.

6. The continuous flow chemical product carbon tracking method based on RTD dynamic compensation according to claim 5, characterized in that, The continuous flow chemical product carbon tracking method based on RTD dynamic compensation also includes feedback correction, that is, using correction gain to compensate for the final carbon footprint, specifically using the following formula: ； ； This represents the final carbon footprint per unit weight of product produced at the current moment after correction. This is the feedback gain coefficient; This indicates the measured concentration at the reactor outlet at the current moment, obtained through sensor measurement. This indicates the model-predicted concentration at the reactor outlet at the current moment; The carbon-matter conversion factor is the conversion coefficient that transforms material concentration deviation into carbon footprint deviation. Indicates the time starting from the current moment The concentration of the material entering the reactor at a time prior to the specified time.

7. The continuous flow chemical product carbon tracking method based on RTD dynamic compensation according to claim 6, characterized in that, The continuous flow chemical product carbon tracking method based on RTD dynamic compensation is subject to the following constraints: Gain saturation limit: The absolute value of the term and An error will be reported if the ratio exceeds the set threshold; Dead zone protection: when If the absolute value of the item is less than the sensor's measurement accuracy limit, no compensation will be performed. Convergence criterion: If the feedback correction direction is opposite for multiple consecutive sampling periods, oscillation occurs, and a forced lock is triggered. It also suggests that manual intervention be performed.

8. The continuous flow chemical product carbon tracking method based on RTD dynamic compensation according to claim 1, characterized in that, The continuous flow chemical product carbon tracking method based on RTD dynamic compensation also includes flow field offset diagnosis, which is used to optimize the model series. The following variance criterion formula is used for flow field migration diagnosis: ； This represents the variance. Indicates the sampling time; Indicates the sampling period; This indicates the average length of stay. , Indicates volume. Indicates flow rate; This is the measured concentration sequence at the reactor outlet; For model series; The model theory has dimensionless variance; Adjust the model level when the preset threshold is exceeded. ,make Reduce to no more than the preset threshold.

9. A continuous flow chemical product carbon tracking system based on RTD dynamic compensation, characterized in that, include: RTD modeling unit is used for RTD mathematical characterization modeling of reactor and generation of probability density sequence: the reactor is abstracted into a single-stage or multi-stage series stirred tank model, and the residence time distribution density function is used to describe the residence probability of fluid particles in the reactor. The carbon load construction unit is used to construct the instantaneous carbon load factor: real-time acquisition of reactor inlet flow, raw material implicit carbon intensity and multiple energy consumption, integrating the scattered energy consumption and material data into a carbon flow signal that varies over time. A convolutional kernel engine for producing carbon based on temporal convolution.

10. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the computer program to implement the steps of the continuous flow chemical product carbon tracking method based on RTD dynamic compensation as described in any one of claims 1-8.

Images