Method for calculating air separation yield by using liquid level of storage tank and change of liquid level

Abstract

The invention provides a method for calculating the air separation yield by using the liquid level of a storage tank and the change of the liquid level. Multi-source data of the storage tank are collected. Liquid level data are collected in real time through a differential pressure liquid level meter, a dynamic liquid level yield mapping algorithm module is started, a correlation model of liquid level change and liquid yield is established through data preprocessing and liquid level variable quantity analysis, and accurate metering of the yield during the period without liquid consumption is achieved. For the intermittent filling process, a multi-component time sequence decomposition algorithm module is introduced, periodic, seasonal and random characteristic decomposition is conducted on liquid level changes, and a correlation model of the liquid level and the filling amount is constructed and used for calculating the single-time filling outflow amount. The two algorithm modules are integrated to a mainboard of the intelligent differential pressure liquid level meter, the yield data fusion algorithm module is started, time alignment and data integration are carried out on non-consumption yield and filling outflow, and the total yield of air separation liquid in any time period is accurately calculated according to the dynamic change rule of the liquid level under different working conditions.

Description

Technical Field

[0001] This invention relates to the field of air separation production calculation, and in particular to a method for calculating air separation production using the liquid level in a storage tank and its changes. Background Technology

[0002] In the scenario of air separation liquid storage and metering, air separation liquid usually flows into liquid storage tanks for storage. Liquid storage tanks are generally only equipped with pressure gauges and differential pressure level gauges that calculate liquid level by pressure difference, and there are no flow meters specifically used to measure air separation output.

[0003] In actual production, liquid storage tanks are filled intermittently and at irregular intervals, causing the liquid level inside the tank to be constantly changing. This dynamic change makes it impossible to accurately capture the actual increase or decrease of air separation liquid in the tank by relying solely on existing differential pressure level gauges. Consequently, it is impossible to effectively measure the air separation output, resulting in missing output data during the air separation production process, which makes it difficult to meet the needs of production management and metering accounting. Summary of the Invention

[0004] The main objective of this invention is to provide a method for calculating air separation production using the liquid level and its changes in a storage tank. This solves the technical problem in the prior art where air separation liquid storage tanks are only equipped with pressure gauges and differential pressure level gauges, and the liquid level changes dynamically due to irregular filling, making it impossible to measure air separation production.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for calculating air separation production using the liquid level in a storage tank and its changes, the method comprising: S1. Identify the air separation liquid storage tank equipped with a differential pressure level gauge and whose liquid level changes dynamically due to intermittent filling as the measurement object. Collect the basic parameters of the storage tank and historical liquid level data of intermittent filling to provide basic data for algorithm modeling. The basic parameters of the storage tank include tank volume and density of the stored liquid medium. S2. Based on the real-time liquid level data collected by the differential pressure level gauge, the dynamic liquid level production mapping algorithm module is started. Through data preprocessing and liquid level change analysis, the correlation between liquid level change and liquid production is established, and the production measurement is completed when there is no liquid consumption. S3. For the liquid outflow process of intermittent filling of storage tanks, start the multi-component time series decomposition algorithm module to decompose the periodic, seasonal and real-time characteristics of liquid level changes, construct the correlation model between liquid level and filling volume, and calculate the outflow volume of a single filling. S4. Integrate the dynamic liquid level production mapping algorithm module of step S2 and the multi-component time-series decomposition algorithm module of step S3 into the main board of the intelligent differential pressure level gauge, start the production data fusion algorithm module, perform time alignment and integration of the production without liquid consumption and the filling outflow data, and finally use the change law of the tank liquid level under different usage conditions to estimate the production of air separation liquid in any time period.

[0006] In the preferred embodiment, the steps of the dynamic liquid level-production mapping algorithm module in step S2 include: S21. Receive the raw liquid level data transmitted in real time from the differential pressure level gauge, and use the sliding window averaging filter method to remove outliers caused by equipment fluctuations. The filtering formula is: ; in, The average liquid level after filtering. This represents the number of sampling points within the window. The original liquid level at the k-th sampling point; S22. Determine the initial liquid level at the start of metering. Current liquid level at the metering point Calculate the change in liquid level ,in, This refers to the liquid level value corresponding to the start time of metering. This represents the liquid level value at the metering point. This measures the change in liquid level over a given time period. S23. Call the tank volume collected in step S1. With medium density Substituting the parameters into the production calculation formula, we obtain the liquid production rate under a scenario with no liquid consumption: ; in, This refers to the liquid production over a period of time when there is no liquid consumption. These are the volume parameters of the storage tank. The density parameter of the medium used to store the air separation liquid.

[0007] In the preferred embodiment, the dynamic liquid level-production mapping algorithm module in step S2 further includes a density-temperature correction sub-step, including: S24. Temperature Acquisition: The temperature of the stored liquid is acquired in real time via a temperature sensor built into the storage tank. ,in, To store the real-time temperature of the air separation liquid; S25. Density Correction: The density of the medium is corrected using a temperature-density correlation formula. ; in, For temperature The actual liquid density below, The standard density of the liquid at 20°C; S26. Corrected Production Calculation: Calculate the corrected output... Substituting into the production formula, we obtain the temperature-compensated production rate without liquid consumption: ; in, This represents the production output after temperature compensation and without liquid consumption.

[0008] In the preferred embodiment, the steps of the multi-component time series decomposition algorithm module in step S3 include: S31. Extract the liquid level data recorded by the differential pressure level gauge during the filling process, and filter out the liquid level at the start of filling. Liquid level at the end of filling and the real-time liquid level sequence during the filling process, among which, The liquid level value at the start of filling. The liquid level value at the moment the filling process ends; S32. Determine the filling cycle by analyzing historical filling data. Based on this periodicity, the periodic variation component is fitted: ; in, These are the values ​​corresponding to the periodically changing components. For the filling time variable, The slope of the periodic change. The intercept is a periodically varying value. S33. Fitting seasonal variation components based on seasonal differences: ; in, These are the numerical values ​​corresponding to the seasonal variation components. The slope of seasonal variation, For seasonal variation intercept; S34. Fitting components that change over time: ; in, To accommodate the changing values ​​of the components, The slope is randomly varying. The intercept is randomly varying.

[0009] In the preferred embodiment, the multi-component time-series decomposition algorithm module in step S3 further includes a sub-step for calculating the filling quantity integration, including: S35. Determine the weights of each component, specifically the weights of the periodic components. Seasonal component weighting Component weights change at any time ,in, This represents the weighting percentage of periodic components in the fill quantity calculation. This represents the weighting of seasonal components in the calculation of fill volume. To ensure that the weighting of components in the filling quantity calculation varies over time; S36. Calculation of single-fill outflow: The single-fill outflow is obtained by summing the weighted sums of the components according to their respective weights. ; in, This refers to the outflow amount per single filling.

[0010] In the preferred embodiment, the multi-component temporal decomposition algorithm module in step S3 further includes a filling mode adaptation sub-step, as detailed below: S37. Determine the filling mode based on historical data of intermittent filling of storage tanks; S38. Adjust the slope of the periodic component for different modes: If it is the first filling mode, adjust the slope of the periodic component to be... Substituting into the formula for the periodic variation component, we obtain the adjusted periodic variation component: ; in, The adjusted periodic variation component, This is the adjusted slope of the periodic change; If it is the second filling mode, adjust the slope of the periodic component to be... Substituting into the formula for the periodic variation component, we obtain the adjusted periodic variation component:

[0011] in, The adjusted periodic variation component, This represents the adjusted slope of the periodic change.

[0012] In the preferred embodiment, the steps of the production data fusion algorithm module in step S4 include: S41. The production data without liquid consumption obtained in step S2 Compared with the filling outflow data obtained in step S3 Convert to timestamp; Among them, production data without liquid consumption Corresponding measurement time period Filling outflow data Corresponding filling time period Then calculate the time overlap interval between the two: ; in, The duration of the time overlap interval. The starting time for measuring production without liquid consumption. The termination time for measuring production without liquid consumption. The start time of a single filling. The end time for a single filling; S42. Calculate the production data without liquid consumption within the overlapping interval in segments to obtain the production output of each segment. Summarize all valid filling outflows within the measurement period. ,in, This refers to segmented production with no liquid consumption. The total amount of effective filling outflow; S43. Sum the segmented production rate with the effective fill outflow rate to obtain the total air separation liquid production rate within any metering time period: ; in, The total output over any given time period. This is the sum of the output of segments with no liquid consumption.

[0013] In the preferred embodiment, the production data fusion algorithm module in step S4 further includes a system error compensation sub-step, including: S44. Periodically calibrate the differential pressure level gauge using a standard flow meter and record the output calculated by the differential pressure level gauge. Production measured by a standard flow meter Calculate the systematic error: ; in, This is the systematic error of the differential pressure level gauge. The output calculated by the differential pressure level gauge. Production volume measured by a standard flow meter; S45. Substitute the systematic error into the total output formula to correct the error: ; in, This represents the final total output after error compensation.

[0014] In the preferred embodiment, the step of collecting historical liquid level data during intermittent filling of the storage tank in step S1 includes: S11. Collect historical liquid level data for a specified duration, where "specified duration" is the time span of historical data that meets the algorithm parameter fitting requirements; S12. Historical data must include the number of fillings, filling start time, filling end time, real-time liquid level during filling, and ambient temperature for the corresponding time period. Among them, "number of fillings" refers to the total number of tank fillings per unit time, "real-time liquid level" refers to the liquid level data collected at fixed intervals during filling, and "ambient temperature" refers to the ambient temperature during the filling time period. S13. Clean the historical data and remove invalid and missing data. "Invalid data" refers to liquid level values ​​that are outside the normal liquid level range of the storage tank, and "missing data" refers to liquid level data that was not collected due to equipment failure.

[0015] In the preferred embodiment, step S4, which involves integrating the algorithm module into the mainboard of the intelligent differential pressure level gauge, includes: S46. Compile the logic code of the dynamic liquid level production mapping algorithm module and the multi-component time series decomposition algorithm module into machine code executable by the microcontroller unit of the intelligent differential pressure liquid level gauge motherboard. Here, the "logic code" is the program code that implements the algorithm function, and the "microcontroller unit" is the core control component of the intelligent differential pressure liquid level gauge motherboard. S47. The compiled machine code is burned into the storage component of the microcontroller unit through the debugging interface, where the "debugging interface" is a hardware interface used for code burning and program debugging, and the "storage component" is a component in the microcontroller unit used to store program code; S48. After the programming is completed, simulate two scenarios: no liquid consumption and intermittent filling. Input test liquid level data to verify the production calculation results output by the motherboard. Among them, "test liquid level data" is the liquid level value generated by simulating the actual working conditions, and "production calculation results" is the production data output by the motherboard according to the algorithm module.

[0016] This invention provides a method for calculating air separation production using the liquid level and its changes in a storage tank. First, it effectively solves the limitation of traditional air separation liquid production measurement relying on flow meters. It eliminates the need for additional specialized flow metering equipment, relying solely on the differential pressure level gauge already installed in the storage tank to perform production calculations. This significantly reduces equipment investment costs and the complexity of installation and commissioning. At the same time, it avoids the impact of flow meter failures or improper maintenance on production measurement, thus improving the economy and stability of the measurement scheme.

[0017] Secondly, this method fully considers the complex operating conditions of air separation liquid storage tanks in actual use, and can accurately adapt to the dynamic changes in liquid level caused by intermittent filling of the storage tank. It combines the liquid level change pattern to calculate the output under different scenarios. It can accurately calculate the liquid output when there is no liquid consumption, and can also accurately measure the liquid outflow during intermittent filling. Thus, it can realize the complete estimation of the total output of air separation liquid in any time period, covering the consumption of liquid storage tanks, ensuring that the output data can fully reflect the actual situation of air separation production and meet the needs of output statistics and accounting at different times in the production process.

[0018] Furthermore, by embedding the core algorithm into the motherboard of the intelligent differential pressure level gauge, the algorithm chip enables real-time data processing and production calculation, eliminating the need for complex manual data organization and calculation. This simplifies the production measurement process, improves measurement efficiency, and allows production measurement to be performed synchronously with level monitoring. This facilitates timely and accurate production data acquisition for staff, providing reliable data support for the scheduling, management, and subsequent production optimization of air separation production, and further ensuring the orderly and efficient operation of the air separation production process. Attached Figure Description

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of the method for calculating air separation output according to the present invention; Figure 2 This is a diagram of the air separation production monitoring system of the present invention; Figure 3 This is a real-time comparison chart of real-time production data and storage tank liquid level data according to the present invention. Detailed Implementation

[0020] Example 1 like Figure 1-3 As shown, a method for calculating air separation production using tank liquid level and its changes is described. The method includes: S1. Identify the air separation liquid storage tank equipped with a differential pressure level gauge and whose liquid level changes dynamically due to intermittent filling as the measurement object. Collect the basic parameters of the storage tank and historical liquid level data of intermittent filling to provide basic data for algorithm modeling. The basic parameters of the storage tank include tank volume and density of the stored liquid medium. S2. Based on the real-time liquid level data collected by the differential pressure level gauge, the dynamic liquid level production mapping algorithm module is started. Through data preprocessing and liquid level change analysis, the correlation between liquid level change and liquid production is established, and the production measurement is completed when there is no liquid consumption. S3. For the liquid outflow process of intermittent filling of storage tanks, start the multi-component time series decomposition algorithm module to decompose the periodic, seasonal and real-time characteristics of liquid level changes, construct the correlation model between liquid level and filling volume, and calculate the outflow volume of a single filling. S4. Integrate the dynamic liquid level production mapping algorithm module of step S2 and the multi-component time-series decomposition algorithm module of step S3 into the main board of the intelligent differential pressure level gauge, start the production data fusion algorithm module, perform time alignment and integration of the production without liquid consumption and the filling outflow data, and finally use the change law of the tank liquid level under different usage conditions to estimate the production of air separation liquid in any time period.

[0021] Step S1 determines the measurement object and basic data, ensuring a reliable data source for subsequent algorithm modeling and avoiding calculation deviations due to missing or inaccurate data. Step S2 activates a dedicated algorithm module for scenarios without liquid consumption, enabling accurate production measurement in this scenario and resolving the issue of unclear correlation between liquid level and production in a single scenario. Step S3 focuses on liquid outflow scenarios, breaking down liquid level change characteristics and constructing a correlation model to overcome the challenge of measuring outflow due to dynamic changes in liquid level caused by intermittent filling. Step S4 integrates the algorithm into the liquid level gauge motherboard and activates the data fusion module, enabling the estimation of production for any time period without a flow meter, directly solving the core problem in the background technology of "intermittent filling of storage tanks and inability to measure production due to the lack of a flow meter".

[0022] The beneficial effects are reflected in three aspects: First, no additional flow meter needs to be installed, reducing equipment investment and maintenance costs. Metering can be achieved by relying on existing differential pressure level gauges, improving the economic efficiency of the solution. Second, scenario-based measurement covers two core operating conditions: no liquid consumption and liquid outflow, ensuring comprehensive production data and meeting the metering needs of different scenarios in production. Third, algorithm integration and data fusion enable real-time metering, simplifying the operation process and providing timely and accurate data support for production scheduling, management, and optimization.

[0023] In the preferred embodiment, the steps of the dynamic liquid level-production mapping algorithm module in step S2 include: S21. Receive the raw liquid level data transmitted in real time from the differential pressure level gauge, and use the sliding window averaging filter method to remove outliers caused by equipment fluctuations. The filtering formula is: ; in, The average liquid level after filtering. This represents the number of sampling points within the window. The original liquid level at the k-th sampling point; S22. Determine the initial liquid level at the start of metering. Current liquid level at the metering point Calculate the change in liquid level ,in, This refers to the liquid level value corresponding to the start time of metering. This represents the liquid level value at the metering point. This measures the change in liquid level over a given time period. S23. Call the tank volume collected in step S1. With medium density Substituting the parameters into the production calculation formula, we obtain the liquid production rate under a scenario with no liquid consumption: ; in, This refers to the liquid production over a period of time when there is no liquid consumption. These are the volume parameters of the storage tank. The density parameter of the medium used to store the air separation liquid.

[0024] The explanation of the S2 step dynamic liquid level production mapping algorithm module aims to clarify the actual operation process of the algorithm module, ensure the feasibility and accuracy of production calculation in scenarios without liquid consumption, and avoid measurement errors caused by ambiguity in the algorithm steps.

[0025] S21 Data Preprocessing: The raw liquid level data is processed by the sliding window averaging filter method, which can effectively remove abnormal values ​​caused by equipment fluctuations (such as instantaneous sensor interference and circuit noise) in the differential pressure level gauge, making the liquid level data more stable and reliable. This lays an accurate data foundation for subsequent liquid level change calculation and production calculation, and improves the accuracy of the entire measurement process.

[0026] S22 Liquid Level Change Calculation: Define the starting point of metering and the current liquid level, calculate the difference between the two to obtain the liquid level change, establish a direct relationship between "time and liquid level change", provide key intermediate parameters for converting liquid level change into production, and ensure that subsequent production calculations have a clear input basis.

[0027] S23 Production Calculation: By calling basic parameters and liquid level changes, production is calculated, directly converting the abstract liquid level change into a specific production value, realizing quantitative measurement of production in scenarios with no liquid consumption, and meeting the needs of production statistics and accounting in this scenario.

[0028] Explanation of Formulas and Letters Filtering formula: ; This represents the average liquid level after filtering. This parameter can eliminate abnormal fluctuations in the original data, making the liquid level data closer to the actual situation. The number of sampling points within the window is used to determine the amount of data involved in the averaging calculation, thereby reducing the impact of a single outlier by averaging multiple sampling points. The original liquid level representing the kth sampling point is the basic input data for the filtering calculation, which comes directly from the real-time acquisition of the differential pressure level gauge.

[0029] The raw liquid level data is smoothed to remove outliers caused by equipment fluctuations, and outputs a stable and accurate average liquid level. This provides high-quality data for subsequent liquid level change calculations, avoiding deviations in liquid level change calculations caused by outliers, which in turn affect the production calculation results.

[0030] Formula for liquid level change: ; It represents the change in liquid level during the measurement period and is the core intermediate parameter connecting liquid level and output. It directly reflects the increase or decrease in the volume of liquid in the storage tank during that period. The liquid level value corresponding to the start time of metering is used as the reference value for calculation to determine the initial state of metering; It represents the liquid level value at the metering point, reflecting the liquid level status of the storage tank at the end of the metering process.

[0031] The liquid levels at the "start" and "end" points are converted into specific differences in value, quantifying the increase or decrease in liquid in the tank. This provides a key variable for subsequent input into the production formula, ensuring that the production calculation has a clear basis for liquid level changes.

[0032] Production calculation formula: ; It represents the liquid production over a period of time when there is no liquid consumption. It is the ultimate target parameter for measurement in this scenario and is directly used for production statistics. The volume parameters of the storage tank reflect its capacity characteristics and determine the proportion of volume change corresponding to the change in liquid level. The density parameter of the medium representing the liquid stored in the air separation process reflects the mass characteristics of the liquid and converts volume changes into mass output. Air separation output is usually measured in mass.

[0033] By combining liquid level change, tank volume, and liquid density, the abstract liquid level change is transformed into a specific mass output, enabling quantitative calculation of output in scenarios with no liquid consumption. This directly meets the core requirements of production measurement in such scenarios, and the calculation logic is simple, highly operable, and easy to integrate with subsequent algorithms.

[0034] In the preferred embodiment, the dynamic liquid level-production mapping algorithm module in step S2 further includes a density-temperature correction sub-step, including: S24. Temperature Acquisition: The temperature of the stored liquid is acquired in real time via a temperature sensor built into the storage tank. ,in, To store the real-time temperature of the air separation liquid; S25. Density Correction: The density of the medium is corrected using a temperature-density correlation formula. ; in, For temperature The actual liquid density below, The standard density of the liquid at 20°C; S26. Corrected Production Calculation: Calculate the corrected output... Substituting into the production formula, we obtain the temperature-compensated production rate without liquid consumption: ; in, This represents the production output after temperature compensation and without liquid consumption.

[0035] The supplementary explanation for step S2, "Dynamic Liquid Level Production Mapping Algorithm Module," adds a density and temperature correction function. This function aims to eliminate the impact of temperature changes on liquid density, further improve the accuracy of production calculation in scenarios without liquid consumption, and avoid density deviations caused by temperature fluctuations, which in turn affect the accuracy of production measurement.

[0036] S24 Temperature Acquisition: The liquid temperature is acquired in real time through the built-in temperature sensor in the storage tank. It can promptly capture temperature changes and provide real-time temperature parameters for density correction, avoiding calculation errors caused by using a fixed density and ensuring that the density parameters are consistent with the actual state of the liquid.

[0037] S25 Density Correction: Based on the temperature-density correlation formula, density is corrected. The density value can be adjusted according to the real-time temperature to eliminate the influence of temperature on liquid density (the density of air separation liquid changes significantly with temperature; as the temperature increases, the density decreases, and vice versa), ensuring the accuracy of the density parameters used for production calculation.

[0038] S26 Corrected Production Calculation: Substitute the corrected density into the production calculation to ensure that the final production result reflects the actual situation after temperature changes, improves the accuracy of production measurement, and is especially suitable for production scenarios with large fluctuations in ambient temperature, ensuring the reliability of production data under different temperature conditions.

[0039] Explanation of Formulas and Letters Density correction formula: ; Represents temperature The actual liquid density shown below is a corrected density parameter used for subsequent accurate production calculations. This represents the standard density of the liquid at 20°C, serving as a reference value for density correction. This value can be obtained through laboratory testing. The real-time temperature of the stored air separation liquid is a key factor causing density changes and directly determines the correction range; 0.0015 is the experimentally determined density temperature coefficient of the air separation liquid, reflecting the proportion of density change corresponding to a unit temperature change.

[0040] The liquid density is dynamically adjusted based on real-time temperature, converting the density at the standard temperature to the density at the actual temperature. This eliminates the influence of temperature fluctuations on density, ensuring that the density parameter used for production calculation matches the actual state of the liquid and avoiding deviations in production calculation due to a fixed density.

[0041] Revised production calculation formula: ; The output with no liquid consumption after temperature compensation is the final output result after temperature correction, and has higher accuracy. With claim 2 The meaning is the same, that is, the change in liquid level within a measurement period; The volume parameter representing the storage tank has the same meaning as in claim 2; The value represents the actual liquid density after modification and is the core parameter that distinguishes this formula from the production formula in claim 2.

[0042] By substituting the temperature-corrected density into the production calculation, a production result that reflects the actual temperature conditions is obtained, further improving the accuracy of production measurement, solving the problem of production calculation deviation caused by temperature fluctuations, and enabling production data to accurately reflect the actual production situation under different temperature environments.

[0043] In the preferred embodiment, the steps of the multi-component time series decomposition algorithm module in step S3 include: S31. Extract the liquid level data recorded by the differential pressure level gauge during the filling process, and filter out the liquid level at the start of filling. Liquid level at the end of filling and the real-time liquid level sequence during the filling process, among which, The liquid level value at the start of filling. The liquid level value at the moment the filling process ends; S32. Determine the filling cycle by analyzing historical filling data. Based on this periodicity, the periodic variation component is fitted: ; in, These are the values ​​corresponding to the periodically changing components. For the filling time variable, The slope of the periodic change. The intercept is a periodically varying value. S33. Fitting seasonal variation components based on seasonal differences: ; in, These are the numerical values ​​corresponding to the seasonal variation components. The slope of seasonal variation, For seasonal variation intercept; S34. Fitting components that change over time: ; in, To accommodate the changing values ​​of the components, The slope is randomly varying. The intercept is randomly varying.

[0044] The explanation of step S3, "Multi-component time-series decomposition algorithm module," is to clarify the operational process of this algorithm module in splitting liquid level change characteristics and constructing a correlation model, ensuring the feasibility and accuracy of calculating the outflow volume of a single filling in the liquid outflow scenario, and overcoming the technical difficulty of measuring the dynamic changes in liquid level during intermittent filling.

[0045] Technical effects and beneficial effects S31 Data Preprocessing: Screening key liquid level data (starting and ending liquid levels and real-time sequences) during the filling process can remove redundant data unrelated to filling, focus on the core information of the filling process, provide accurate data input for subsequent analysis of liquid level change characteristics, and avoid redundant data interfering with calculations.

[0046] S32 Periodic Component Extraction: By statistically analyzing historical data to determine the filling cycle and fitting periodic components, it can capture recurring regular changes during the filling process (such as filling at fixed times every day, or filling for similar durations each time), separating periodic changes from complex liquid level changes, and providing a regular basis for accurate measurement of outflow.

[0047] S33 Seasonal component extraction: Based on seasonal differences, seasonal components are fitted, which can take into account the impact of seasonal environmental changes on liquid level (such as rapid liquid evaporation in summer leading to a rapid drop in liquid level, and slow evaporation in winter leading to a slow drop in liquid level), eliminating the interference of seasonal factors on outflow calculation and improving the accuracy of measurement in different seasons.

[0048] S34 Real-time Component Extraction: Fits real-time components to capture liquid level fluctuations caused by random factors such as equipment failure and operational deviations. Separates random changes from periodic and seasonal variations, avoiding random factors from masking normal liquid level change patterns and ensuring that outflow calculations reflect the true filling situation.

[0049] Explanation of Formulas and Letters Formula for periodic components: ; The numerical value representing the periodic change component reflects the part of the liquid level change that repeats with a fixed period and is the core component of the liquid level change during the filling process. Representing the filling time variable, it reflects the time progress of the filling process and is a periodically changing related variable; It represents the slope of the periodic change, reflecting the rate of change of the periodic change over time (such as the magnitude of the drop in liquid level per unit time due to periodic factors). It represents the intercept of periodic changes and reflects the initial reference value of periodic changes (such as the initial liquid level deviation corresponding to periodic factors at the start of filling).

[0050] By fitting the periodic component of liquid level changes with a linear function, the periodicity is quantified, making it easier to separate the periodic influence from complex liquid level changes. This provides regular component data for subsequent integrated calculation of outflow, ensuring accurate capture of the periodic characteristics of the filling process.

[0051] Seasonal component formula: ; The numerical value representing the seasonal variation component reflects the portion of liquid level changes caused by seasonal differences and is an important component of liquid level changes during the filling process. This represents the filling time variable, which has the same meaning as in the periodic component formula and is a related variable that varies seasonally. It represents the slope of seasonal variation, reflecting the rate of change of seasonal changes over time (e.g., the drop in liquid level per unit time is greater in summer than in winter due to seasonal factors). It represents the intercept of seasonal variation and reflects the initial baseline value of seasonal variation (such as the initial liquid level deviation corresponding to seasonal factors at the start of filling in a certain season).

[0052] By fitting the seasonal component of liquid level changes with a linear function, the seasonal impact is quantified, eliminating the interference of seasonal factors on the outflow calculation. This ensures that the outflow calculation in different seasons can accurately reflect the actual filling situation and avoid measurement deviations caused by seasonal differences.

[0053] Formula for components that change over time: ; It represents the numerical value corresponding to the variable component at any time, reflecting the part of the liquid level change caused by random factors, and is a supplementary component of the liquid level change in the filling process. This represents the filling time variable, which has the same meaning as in the previous two component formulas; it is a variable that changes over time. Represents the slope of random changes, reflecting the rate of change of liquid level over time caused by random factors, including the magnitude of abnormal liquid level changes per unit time caused by instantaneous equipment failure. Represents the intercept of random variation, reflecting the initial liquid level deviation corresponding to random factors, including, for example, initial liquid level anomalies caused by operational deviations at the start of filling.

[0054] By fitting the random component of liquid level changes with a linear function, the random influence is quantified, avoiding random factors from masking the normal liquid level change pattern. This ensures that the outflow calculation can eliminate random interference and accurately reflect the true situation of the filling process.

[0055] In the preferred embodiment, the multi-component time-series decomposition algorithm module in step S3 further includes a sub-step for calculating the filling quantity integration, including: S35. Determine the weights of each component, specifically the weights of the periodic components. Seasonal component weighting Component weights change at any time ,in, This represents the weighting percentage of periodic components in the fill quantity calculation. This represents the weighting of seasonal components in the calculation of fill volume. To ensure that the weighting of components in the filling quantity calculation varies over time; S36. Calculation of single-fill outflow: The single-fill outflow is obtained by summing the weighted sums of the components according to their respective weights. ; in, This refers to the outflow amount per single filling.

[0056] The explanation of step S3, "Multi-component time-series decomposition algorithm module," clarifies the filling volume integration calculation process. Its function is to integrate the three liquid level change components after decomposition into a single filling outflow volume, thereby realizing the quantitative measurement of filling volume in liquid outflow scenarios and solving the problem of "how to integrate multiple components to obtain the final outflow volume."

[0057] S35 Component Weight Determination: Clearly define the weight ratio of each component. The weight can be allocated according to the degree of influence of each component on the filling volume (periodic components have the greatest influence and the highest weight; components that change at any time have the least influence and the lowest weight). This ensures that the core influencing factors are highlighted during the integrated calculation, avoids secondary factors from interfering with the final result, and improves the rationality of the outflow calculation.

[0058] S36 Single Filling Outflow Calculation: By weighted summation and integration of various components, the multi-dimensional liquid level change components are transformed into a single outflow value, realizing quantitative measurement of filling volume, directly meeting the measurement needs in liquid outflow scenarios, and the weighted integration method can comprehensively consider the influence of different factors to ensure accurate and reliable outflow data.

[0059] Explanation of Formulas and Letters Formula for single-fill outflow: ; Representing the outflow volume of a single filling, it is the core target parameter for measurement in liquid outflow scenarios and is directly used to count the liquid outflow volume of each filling. This represents the weighting percentage of the periodic component in the filling volume calculation, reflecting the degree of influence of the periodic component on the outflow. The weighting value is determined based on the actual magnitude of the influence. The values ​​representing the periodically changing components are consistent with the meanings in S31-S34; This represents the weighting of seasonal components in the filling volume calculation, reflecting the degree of influence of seasonal components on outflow. The values ​​representing the seasonal variation components are consistent with the meanings in S31-S34; This represents the weighting of the variable component in the filling volume calculation, reflecting the degree of influence of the variable component on the outflow. This represents the numerical value corresponding to the component that changes over time, and has the same meaning as S31-S34.

[0060] The three components of cycle, season, and randomness are integrated according to their weights, and the single filling outflow is obtained by taking into account the influence of various factors. This achieves accurate measurement of filling volume under the influence of multiple factors, avoids the deviation caused by calculation of a single factor, and ensures that the outflow data can fully reflect the actual situation of the filling process, providing accurate single outflow data for subsequent total production integration.

[0061] In the preferred embodiment, the multi-component temporal decomposition algorithm module in step S3 further includes a filling mode adaptation sub-step, as detailed below: S37. Determine the filling mode based on historical data of intermittent filling of storage tanks; S38. Adjust the slope of the periodic component for different modes: If it is the first filling mode, adjust the slope of the periodic component to be... Substituting into the formula for the periodic variation component, we obtain the adjusted periodic variation component: ; in, The adjusted periodic variation component, This is the adjusted slope of the periodic change; If it is the second filling mode, adjust the slope of the periodic component to be... Substituting into the formula for the periodic variation component, we obtain the adjusted periodic variation component:

[0062] in, The adjusted periodic variation component, This represents the adjusted slope of the periodic change.

[0063] Further explanation of step S3, "Multi-component time-series decomposition algorithm module," adds a filling mode adaptation function. Its function is to adjust the periodic component parameters according to different filling modes, ensuring that the periodic components can still accurately reflect the liquid level change pattern under different filling frequencies and durations, thereby improving the accuracy of outflow calculation under different filling modes.

[0064] S37 Filling Mode Determination: Based on historical data, the filling mode is determined, which can identify different filling patterns in production (such as high frequency and short duration, low frequency and long duration), providing a basis for subsequent parameter adjustment and avoiding calculation deviations caused by using the same parameters for different modes.

[0065] S38 Mode Parameter Adjustment: Adjusting the slope of the periodic component for different modes allows the periodic component to adapt to the liquid level change rate under different modes (in high-frequency, short-duration modes, the filling interval is short, the liquid residence time is short, and the liquid level change rate is fast, so the slope needs to be increased; in low-frequency, long-duration modes, the opposite is true), ensuring that the periodic component can accurately reflect the liquid level change pattern under different modes, thereby improving the accuracy of outflow calculation, making the method applicable to various filling conditions, and enhancing the applicability of the scheme.

[0066] Formula for periodic components in the first filling mode: ; This represents the periodic variation component after adjustment under the first filling mode. It is a periodic parameter adapted to this mode and is used to accurately calculate the outflow under this mode. This represents the adjusted slope of the cycle change in the first filling mode, which is 1.2 times the original slope, to suit the fast liquid level change rate characteristic of this mode. This represents the filling time variable, consistent with the meaning in S31-S34; The intercept represents the periodic change, which has the same meaning as in S31-S34. Since the intercept mainly reflects the initial deviation and is not affected by the filling mode, it is not adjusted.

[0067] By adjusting the slope, the periodic component is adapted to the liquid level change rate of the first filling mode (high frequency, short duration), ensuring that the periodic component can accurately reflect the periodic change pattern of the liquid level in this mode, and avoiding deviation of the periodic component due to a fixed slope, which in turn affects the outflow calculation results.

[0068] The formula for the periodic components of the second filling mode:

[0069] The periodic variation component representing the adjustment under the second filling mode is a periodic parameter adapted to this mode, used to accurately calculate the outflow under this mode. This represents the adjusted slope of the cycle change in the second filling mode, which is 0.8 times the original slope, to suit the slow rate of liquid level change in this mode. This represents the filling time variable, consistent with the meaning in S31-S34; This represents the intercept of the periodic change, which has the same meaning as in S31-S34, and is not adjusted.

[0070] By adjusting the slope to adapt the periodic component to the liquid level change rate of the second filling mode (low frequency, long duration), the periodic component can accurately reflect the periodic change law of the liquid level in this mode, further improving the accuracy of outflow calculation under different filling modes and enhancing the applicability of the method.

[0071] In the preferred embodiment, the steps of the production data fusion algorithm module in step S4 include: S41. The production data without liquid consumption obtained in step S2 Compared with the filling outflow data obtained in step S3 Convert to timestamp; Among them, production data without liquid consumption Corresponding measurement time period Filling outflow data Corresponding filling time period Then calculate the time overlap interval between the two: ; in, The duration of the time overlap interval. The starting time for measuring production without liquid consumption. The termination time for measuring production without liquid consumption. The start time of a single filling. The end time for a single filling; S42. Calculate the production data without liquid consumption within the overlapping interval in segments to obtain the production output of each segment. Summarize all valid filling outflows within the measurement period. ,in, This refers to segmented production with no liquid consumption. The total amount of effective filling outflow; S43. Sum the segmented production rate with the effective fill outflow rate to obtain the total air separation liquid production rate within any metering time period: ; in, The total output over any given time period. This is the sum of the output of segments with no liquid consumption.

[0072] The explanation of step S4, "Production Data Fusion Algorithm Module," clarifies the operational process of total production integration. Its function is to integrate the production data without liquid consumption with the filling outflow data to obtain the total production for any time period. This solves the problem of "how to integrate the data after calculating by scenario to obtain the final total production," and realizes the leap from scenario-based measurement to overall measurement.

[0073] S41 Timeline Alignment: Calculates the time overlap between production output without liquid consumption and filling outflow, ensuring that only data within the same time period is integrated. This avoids duplicate or missed production calculations due to inconsistent time ranges, ensuring the accuracy of the time dimension in total production calculation.

[0074] S42 Data Integration: Calculates and summarizes the effective filling outflow for segments of production without liquid consumption. It can break down production without liquid consumption by filling time period, making the time granularity of the two consistent. At the same time, it filters out the effective outflow that overlaps with the metering time period, providing matching data for subsequent summation and avoiding calculation errors caused by data dimension mismatch.

[0075] S43 Total Production Calculation: The total production for any time period is obtained by summing, directly achieving the core objective of "using liquid level changes to estimate the production for any time period". This meets the production needs for production statistics for different time periods (such as hours, days, and months), and the integration method is logically clear, ensuring that the total production data is accurate and complete.

[0076] Formula for time overlap intervals: ; The time overlap interval between the production volume without liquid consumption and the outflow volume of a single filling is a key parameter for determining whether the two are within the same time period. If the duration is greater than 0, it indicates that there is overlap and the data can be integrated. This represents the start time for measuring the production volume without liquid consumption, and determines the starting point of the production volume without liquid consumption. This represents the end time for measuring the production volume without liquid consumption, and determines the time endpoint for measuring the production volume without liquid consumption. Represents the start time of a single filling, determining the starting point of the outflow amount for a single filling; This represents the end time of a single filling, determining the time endpoint for the outflow of a single filling volume.

[0077] Formula function: By calculating the time overlap interval, it filters out the filling outflow data that overlaps with the time period of production without liquid consumption, avoiding the inclusion of outflows that are not in the same time period in the calculation, and ensuring that the time range of the total production calculation is accurate, without duplication or omission.

[0078] Total output formula: ; Representing the total output over any time period, it is the final output parameter of the entire method and is directly used for total production statistics and accounting. This represents the sum of the output of the segments with no liquid consumption, which is the cumulative value of the output of each segment after the output of no liquid consumption is divided according to the filling time period, ensuring that it is consistent with the time granularity of the filling outflow. It represents the sum of the effective filling outflow, that is, the cumulative value of all single filling outflows that overlap with the metering time period.

[0079] The total output for any time period is obtained by summing the time-aligned, liquid-free segmented output with the effective filling outflow, thus realizing the complete process of "scenario-based measurement - time alignment - total output integration". This directly meets the production needs for output measurement in any time period and ensures that the total output data is comprehensive and accurate.

[0080] In the preferred embodiment, the production data fusion algorithm module in step S4 further includes a system error compensation sub-step, including: S44. Periodically calibrate the differential pressure level gauge using a standard flow meter and record the output calculated by the differential pressure level gauge. Production measured by a standard flow meter Calculate the systematic error: ; in, This is the systematic error of the differential pressure level gauge. The output calculated by the differential pressure level gauge. Production volume measured by a standard flow meter; S45. Substitute the systematic error into the total output formula to correct the error: ; in, This represents the final total output after error compensation.

[0081] The S4 step, "Production Data Fusion Algorithm Module," is supplemented by adding a system error compensation function. Its purpose is to eliminate the impact of the differential pressure level gauge's own accuracy deviation on the total production, further improve the accuracy of total production measurement, and solve the problem of "production calculation deviation caused by the level gauge's own error."

[0082] S44 Error Calibration: Regularly calibrating the level gauge using a standard flow meter allows for the acquisition of the real-time system error of the level gauge. This prevents error accumulation due to decreased accuracy caused by long-term use of the level gauge, ensuring accurate and real-time error parameters and providing a reliable basis for subsequent compensation.

[0083] S45 Error Compensation: Substituting system errors into the total output formula for correction can eliminate the influence of the level gauge's own deviation on the output (such as the level gauge's measured value being generally too high or too low, which can be adjusted to the actual value through error compensation), making the final total output result closer to the real output, improving measurement accuracy, and meeting the needs of high-precision production measurement.

[0084] Systematic error formula: ; The systematic error of the differential pressure level gauge is expressed as a percentage and reflects the degree of deviation between the measured value and the true value. It is the core parameter for error compensation. This represents the output calculated by the differential pressure level gauge, i.e., the output obtained by this method before error compensation. The output measured by the standard flow meter is represented by the standard flow meter. The standard flow meter has high accuracy, and its measured value serves as the benchmark for the actual output. The absolute value sign ensures that the error is positive. The numerator represents the absolute deviation between the liquid level gauge's calculated value and the standard value, while the denominator represents the standard value, so that the error is presented in the form of a relative value, which better reflects the degree of deviation.

[0085] By comparing with a standard flow meter, the systematic error of the differential pressure level gauge is calculated, providing a quantifiable deviation parameter for subsequent error compensation. This avoids the lack of a basis for compensation due to the inability to quantify the level gauge error, and ensures the accuracy of error compensation.

[0086] Formula for total output after error compensation: ; The final total output after error compensation is the final output parameter after precision correction, which has higher accuracy. This represents the systematic error of the differential pressure level gauge, and its meaning is consistent with that in the error calibration formula. Represents the total output without error compensation, as shown in the total output formula of claim 7. The meanings are the same.

[0087] By multiplying by a coefficient of (1 - error), the uncompensated total output is corrected to the output after eliminating the system error of the level gauge, making the final result closer to the actual output, further improving the accuracy of total output measurement, meeting the needs of high-precision measurement in production, and is especially suitable for air separation production scenarios with high requirements for output accuracy.

[0088] In the preferred embodiment, the step of collecting historical liquid level data during intermittent filling of the storage tank in step S1 includes: S11. Collect historical liquid level data for a specified duration, where "specified duration" is the time span of historical data that meets the algorithm parameter fitting requirements; S12. Historical data must include the number of fillings, filling start time, filling end time, real-time liquid level during filling, and ambient temperature for the corresponding time period. Among them, "number of fillings" refers to the total number of tank fillings per unit time, "real-time liquid level" refers to the liquid level data collected at fixed intervals during filling, and "ambient temperature" refers to the ambient temperature during the filling time period. S13. Clean the historical data and remove invalid and missing data. "Invalid data" refers to liquid level values ​​that are outside the normal liquid level range of the storage tank, and "missing data" refers to liquid level data that was not collected due to equipment failure.

[0089] The explanation of step S1, "Collecting historical liquid level data of intermittent filling of storage tanks," clarifies the requirements and procedures for data collection. Its purpose is to ensure the reliability of historical liquid level data, meet the needs of algorithm modeling, and avoid subsequent algorithm parameter fitting deviations due to non-compliant data, thus laying a data foundation for the smooth implementation of the entire method.

[0090] Technical effects and beneficial effects S11 Data Duration: The duration of historical data collected to meet the requirements of algorithm parameter fitting ensures that the data covers a sufficient number of filling scenarios (such as different dates and different production loads), making the parameters fitted by the algorithm more representative and avoiding one-sided parameter fitting due to insufficient data duration, which would affect the accuracy of subsequent calculations.

[0091] S12 Data Integrity: Clearly define the core information that historical data should include, ensuring comprehensive data dimensions and covering key parameters of the filling process (number of times, time, liquid level, temperature). This provides sufficient information support for the algorithm to decompose liquid level change characteristics (such as periodic and seasonal changes), avoiding the inability of the algorithm to accurately decompose features due to missing data dimensions.

[0092] S13 Data Preprocessing: Cleaning the data removes invalid and missing data, eliminating "noise" in the data (such as outliers exceeding normal liquid levels or missing values ​​caused by equipment malfunctions), making the data purer and more reliable, avoiding invalid data from interfering with algorithm modeling and parameter fitting, and ensuring high-quality data input for subsequent calculations.

[0093] Specified duration: refers to the historical data time span required to meet the algorithm parameter fitting requirements. It usually needs to cover at least 3 months to ensure that it can include filling data under different seasons and production loads, so that the periodic and seasonal component parameters of the algorithm fitting are more accurate.

[0094] Number of fillings: refers to the total number of times the storage tank is filled per unit time, such as the number of fillings per day or per week. It is a key basis for judging the filling mode (high frequency, low frequency) and provides data support for the determination of the filling mode in claim 6.

[0095] Real-time liquid level value: refers to the liquid level data collected at fixed intervals during the filling process. The interval is usually consistent with the sampling period of the differential pressure level gauge (1-5 minutes / time). It is the basic data for the algorithm to decompose the liquid level change characteristics and ensure that the continuous change of liquid level during the filling process can be captured.

[0096] Ambient temperature: refers to the external ambient temperature during the filling period. It is a key factor affecting the seasonal changes in liquid level and provides temperature-related data for the extraction of seasonal components in claim 4, ensuring that the seasonal components can accurately reflect the influence of ambient temperature.

[0097] Invalid data: refers to liquid level values ​​that exceed the normal range of the storage tank, such as liquid level below the reasonable range of the bottom of the storage tank or above the reasonable range of the top. Such data is usually caused by sensor failure or data transmission error. If it is not removed, it will interfere with the judgment of the liquid level change pattern.

[0098] Missing data: refers to liquid level data that was not collected due to equipment failure (such as temporary malfunction of the level gauge or interruption of data transmission). If it is not removed, the liquid level change sequence will be discontinuous, affecting the algorithm's decomposition and fitting of liquid level change features.

[0099] In the preferred embodiment, step S4, which involves integrating the algorithm module into the mainboard of the intelligent differential pressure level gauge, includes: S46. Compile the logic code of the dynamic liquid level production mapping algorithm module and the multi-component time series decomposition algorithm module into machine code executable by the microcontroller unit of the intelligent differential pressure liquid level gauge motherboard. Here, the "logic code" is the program code that implements the algorithm function, and the "microcontroller unit" is the core control component of the intelligent differential pressure liquid level gauge motherboard. S47. The compiled machine code is burned into the storage component of the microcontroller unit through the debugging interface, where the "debugging interface" is a hardware interface used for code burning and program debugging, and the "storage component" is a component in the microcontroller unit used to store program code; S48. After the programming is completed, simulate two scenarios: no liquid consumption and intermittent filling. Input test liquid level data to verify the production calculation results output by the motherboard. Among them, "test liquid level data" is the liquid level value generated by simulating the actual working conditions, and "production calculation results" is the production data output by the motherboard according to the algorithm module.

[0100] The explanation of step S4, "Integrating the algorithm module into the motherboard of the intelligent differential pressure level gauge," clarifies the implementation method of algorithm integration. Its purpose is to transform the abstract algorithm module into a hardware-executable function, ensuring that the algorithm can run stably on the actual level gauge hardware, realizing the "algorithm-hardware" implementation, and solving the problem of "how to implement the algorithm on the existing level gauge."

[0101] S46 Algorithm Compilation: Compiles the logic code into machine code executable by the microcontroller unit of the level gauge motherboard. This enables the algorithm to adapt to the instruction system of the level gauge hardware, ensuring that the algorithm can run on the level gauge. It realizes the transformation of the algorithm from software logic to hardware executable code, laying the foundation for subsequent programming and operation.

[0102] S47 code burning: By burning the machine code to the storage component of the microcontroller through the debugging interface, the algorithm code can be solidified into the level gauge hardware, ensuring that the algorithm can be automatically loaded and run after the level gauge is started, realizing the long-term stable operation of the algorithm and avoiding the trouble of reloading the algorithm every time it is used.

[0103] S48 Functional Testing: Simulated scenarios verify the production calculation results, ensuring that the level gauge functions normally after algorithm integration and the calculation results are accurate. This avoids the algorithm failing to run properly due to compilation errors, burning failures, or hardware incompatibility, ensuring that the entire method can be reliably implemented on actual hardware.

[0104] Logic code: refers to the program code that implements the algorithm's functions. It is usually written in C language and contains the logic of all algorithm steps in claims 2-9 (such as data filtering, component splitting, yield calculation, error compensation, etc.). It is the software implementation of the algorithm.

[0105] Microcontroller Unit (MCU): Model STM32F103, is the core control component of the mainboard of the intelligent differential pressure level gauge. It is responsible for executing algorithm code, processing level data, calculating output, and outputting results. It is equivalent to the "brain" of the level gauge, and its performance directly determines the efficiency and stability of the algorithm operation.

[0106] Machine code: refers to the binary instructions that the microcontroller unit can directly recognize and execute. It is generated by compiling logic code and is the final form of the algorithm running on hardware, ensuring that the microcontroller unit can understand and execute the algorithm logic.

[0107] Debugging interface: refers to the hardware interface used for code burning and program debugging, usually the JTAG interface. Through this interface, the compiled machine code can be transferred to the storage component of the microcontroller unit. It can also be used to debug problems in the algorithm operation to ensure successful code burning and normal algorithm operation.

[0108] Storage component: refers to the component in the microcontroller unit used to store program code, usually flash memory. Its capacity must be sufficient to hold the algorithm code, and it must also be non-volatile (data is not lost after power failure) to ensure that the algorithm code can be stored in the level gauge for a long time and can be loaded normally each time it is started.

[0109] Test liquid level data: refers to the liquid level values ​​generated under simulated actual working conditions, covering liquid level changes under two scenarios: no liquid consumption and intermittent filling (such as a steady rise in liquid level when simulating no consumption and a rapid drop in liquid level when simulating filling), used to verify whether the output calculated by the liquid level gauge after algorithm integration is accurate.

[0110] Production calculation result: This refers to the production data output by the algorithm module based on the motherboard. That is, after inputting the test liquid level data, the liquid level gauge obtains the production value by running the algorithm. By comparing this result with the preset theoretical production, it can be determined whether the algorithm integration is successful and whether the calculation is accurate.

[0111] Example 2 Further explanation in conjunction with Example 1, such as Figure 1-3 As shown, when actually deploying this "method for calculating air separation production using tank liquid level and its changes", the following steps must be followed, along with the corresponding software and modules: First, preliminary preparations must be made, confirming that the air separation liquid storage tank is equipped with a differential pressure level gauge (intelligent differential pressure level gauge, model Rosemount 3051CD or Siemens SITRANS P series), and that the tank has a built-in temperature sensor (such as a PT100 platinum resistance temperature sensor) to collect the real-time liquid temperature; then, the basic parameters of the storage tank are collected, and the tank volume is determined through the tank design drawings. The standard density of the space-time liquid at 20℃ was obtained through laboratory testing. Meanwhile, historical liquid level data of intermittent filling of storage tanks over the past 3 months (including the number of fillings per day, the start / end time of each filling, the liquid level value every 1-5 minutes during the filling process, and the corresponding ambient temperature) were collected. The historical data was preprocessed using Excel or Python's Pandas library to remove invalid data that exceeds the normal liquid level range of the storage tank and missing data due to equipment failure, ensuring data integrity and reliability.

[0112] Next, the algorithm code is developed and compiled. The logic code for the "Dynamic Liquid Level-Production Mapping Algorithm Module", "Multi-Component Time Series Decomposition Algorithm Module", and "Production Data Fusion Algorithm Module" is written in C language. The core algorithm logic such as sliding window average filtering, autocorrelation function period analysis, and analytic hierarchy process weight calculation needs to be embedded in the code. After the code is written, embedded development software such as Keil MDK (for STM32 series MCUs) or IAR Embedded Workbench is used to compile the C language code into machine code that can be executed by the intelligent differential pressure level gauge motherboard microcontroller unit (MCU, model STM32F103). During the compilation process, it is necessary to ensure that the code occupies no more than 40% of the total capacity of the MCU flash memory, and reserve space for subsequent algorithm upgrades.

[0113] Next, the algorithm module is programmed and integrated into the hardware. The compiled machine code is programmed into the MCU flash memory of the intelligent differential pressure level gauge motherboard through the JTAG debugging interface. The programming tool can be ST-Link V2. After programming, the signal line of the temperature sensor built into the tank is connected to the analog input interface (such as the 4-20mA signal interface) of the level gauge motherboard to ensure that the temperature data can be transmitted to the MCU in real time. At the same time, check whether the liquid level data acquisition function of the differential pressure level gauge is normal to ensure that the liquid level data and temperature data can be transmitted to the algorithm module synchronously.

[0114] Then, functional testing and parameter calibration are performed, simulating a scenario with no liquid consumption, by inputting preset liquid level change data (such as from...) into the level gauge. Gradually rise to The output of the "Dynamic Liquid Level-Production Mapping Algorithm Module" can be viewed through the display interface of the liquid level gauge or through host computer software (such as KingSCADA or WinCC). Verify whether the calculation results match the theoretical values ​​( The data must be consistent with the output of the multi-component time-series decomposition algorithm module, with an error ≤2%. Then, simulate intermittent filling scenarios, inputting liquid level data and ambient temperature data under different filling modes (high-frequency short duration, low-frequency long duration) to verify the single-fill outflow rate output by the module. For accuracy, the error must be ≤3%; finally, the system error is calibrated using a standard flow meter, and the output calculated by the level gauge is recorded synchronously within the same time period. Production measured by a standard flow meter Calculate systematic error using Excel and will Enter the parameters into the "Production Data Fusion Algorithm Module" and complete the error compensation settings.

[0115] Finally, actual operation and data monitoring were conducted. The debugged intelligent differential pressure level gauge was installed back into the storage tank. After the level gauge was started, the algorithm module automatically collected level and temperature data in real time. The "Dynamic Level-Production Mapping Algorithm Module" calculated the production output without liquid consumption, the "Multi-Component Time Series Decomposition Algorithm Module" calculated the single-fill outflow, and the "Production Data Fusion Algorithm Module" completed time alignment, data integration, and error compensation. Finally, the total production output for any time period was output. Staff can view real-time production data through the local display interface of the level gauge, or remotely monitor production data through host computer software (such as KingSCADA or WinCC). They should also periodically (once a month) repeat the system error calibration steps and update the data. These parameters ensure the accuracy of output measurement during long-term operation.

[0116] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be limited to the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A method for calculating air separation production using tank liquid level and its changes, characterized in that: The method includes: S1. Identify the air separation liquid storage tank equipped with a differential pressure level gauge and whose liquid level changes dynamically due to intermittent filling as the measurement object. Collect the basic parameters of the storage tank and historical liquid level data of intermittent filling to provide basic data for algorithm modeling. The basic parameters of the storage tank include tank volume and density of the stored liquid medium. S2. Based on the real-time liquid level data collected by the differential pressure level gauge, the dynamic liquid level production mapping algorithm module is started. Through data preprocessing and liquid level change analysis, the correlation between liquid level change and liquid production is established, and the production measurement is completed when there is no liquid consumption. S3. For the liquid outflow process of intermittent filling of storage tanks, start the multi-component time series decomposition algorithm module to decompose the periodic, seasonal and real-time characteristics of liquid level changes, construct the correlation model between liquid level and filling volume, and calculate the outflow volume of a single filling. S4. Integrate the dynamic liquid level production mapping algorithm module of step S2 and the multi-component time-series decomposition algorithm module of step S3 into the main board of the intelligent differential pressure level gauge, start the production data fusion algorithm module, perform time alignment and integration of the production without liquid consumption and the filling outflow data, and finally use the change law of the tank liquid level under different usage conditions to estimate the production of air separation liquid in any time period.

2. The method for calculating air separation production using tank liquid level and its changes, as described in claim 1, is characterized in that: The steps of the dynamic liquid level production mapping algorithm module in step S2 include: S21. Receive the raw liquid level data transmitted in real time from the differential pressure level gauge, and use the sliding window averaging filter method to remove outliers caused by equipment fluctuations. The filtering formula is: ； in, The average liquid level after filtering. This represents the number of sampling points within the window. The original liquid level at the k-th sampling point; S22. Determine the initial liquid level at the start of metering. Current liquid level at the metering point Calculate the change in liquid level ,in, This refers to the liquid level value corresponding to the start time of metering. This represents the liquid level value at the metering point. This measures the change in liquid level over a given time period. S23. Call the tank volume collected in step S1. With medium density Substituting the parameters into the production calculation formula, we obtain the liquid production rate under a scenario with no liquid consumption: ； in, This refers to the liquid production over a period of time when there is no liquid consumption. These are the volume parameters of the storage tank. The density parameter of the medium used to store the air separation liquid.

3. The method for calculating air separation production using tank liquid level and its changes, as described in claim 2, is characterized in that: The dynamic liquid level and production mapping algorithm module in step S2 also includes a density and temperature correction sub-step, including: S24. Temperature Acquisition: The temperature of the stored liquid is acquired in real time via a temperature sensor built into the storage tank. ,in, To store the real-time temperature of the air separation liquid; S25. Density Correction: The density of the medium is corrected using a temperature-density correlation formula. ； in, For temperature The actual liquid density below, The standard density of the liquid at 20°C; S26. Corrected Production Calculation: Calculate the corrected output... Substituting into the production formula, we obtain the temperature-compensated production rate without liquid consumption: ； in, This represents the production output after temperature compensation and without liquid consumption.

4. The method for calculating air separation production using tank liquid level and its changes, as described in claim 1, is characterized in that: The steps of the multi-component time series decomposition algorithm module in step S3 include: S31. Extract the liquid level data recorded by the differential pressure level gauge during the filling process, and filter out the liquid level at the start of filling. Liquid level at the end of filling and the real-time liquid level sequence during the filling process, among which, The liquid level value at the start of filling. The liquid level value at the moment the filling process ends; S32. Determine the filling cycle by analyzing historical filling data. Based on this periodicity, the periodic variation component is fitted: ； in, These are the numerical values ​​corresponding to the periodically changing components. For the filling time variable, The slope of the periodic change. The intercept is a periodically varying value. S33. Fitting seasonal variation components based on seasonal differences: ； in, These are the numerical values ​​corresponding to the seasonal variation components. The slope of seasonal variation, For seasonal variation intercept; S34. Fitting components that change over time: ； in, To accommodate the changing values ​​of the components, The slope is randomly varying. The intercept is randomly varying.

5. The method for calculating air separation production using the tank liquid level and its changes according to claim 4, characterized in that: The multi-component time series decomposition algorithm module in step S3 also includes a sub-step for filling quantity integration calculation, including: S35. Determine the weights of each component, specifically the weights of the periodic components. Seasonal component weighting Component weights change at any time ,in, This represents the weighting percentage of periodic components in the fill quantity calculation. This represents the weighting of seasonal components in the calculation of fill volume. To ensure that the weighting of components in the filling quantity calculation varies over time; S36. Calculation of single-fill outflow: The single-fill outflow is obtained by summing the weighted sums of the components according to their respective weights. ； in, This refers to the outflow amount per single filling.

6. The method for calculating air separation production using tank liquid level and its changes according to claim 1, characterized in that: The multi-component temporal decomposition algorithm module in step S3 also includes a filling mode adaptation sub-step, as follows: S37. Determine the filling mode based on historical data of intermittent filling of storage tanks; S38. Adjust the slope of the periodic component for different modes: If it is the first filling mode, adjust the slope of the periodic component to be... Substituting into the formula for the periodic variation component, we obtain the adjusted periodic variation component: ； in, The adjusted periodic variation component, This is the adjusted slope of the periodic change; If it is the second filling mode, adjust the slope of the periodic component to be... Substituting into the formula for the periodic variation component, we obtain the adjusted periodic variation component: in, The adjusted periodic variation component, This represents the adjusted slope of the periodic change.

7. The method for calculating air separation production using tank liquid level and its changes, as described in claim 1, is characterized in that: The steps in the production data fusion algorithm module of step S4 include: S41. The production data without liquid consumption obtained in step S2 Compared with the filling outflow data obtained in step S3 Convert to timestamp; Among them, production data without liquid consumption Corresponding measurement time period Filling outflow data Corresponding filling time period Then calculate the time overlap interval between the two: ； in, The duration of the time overlap interval. The starting time for measuring production without liquid consumption. The termination time for measuring production without liquid consumption. The start time of a single filling. The end time for a single filling; S42. Calculate the production data without liquid consumption within the overlapping interval in segments to obtain the production output of each segment. Summarize all valid filling outflows within the measurement period. ,in, This refers to segmented production with no liquid consumption. The total effective filling outflow; S43. Sum the segmented production rate with the effective fill outflow rate to obtain the total air separation liquid production rate within any metering time period: ； in, The total output over any given time period. This is the sum of the output of segments with no liquid consumption.

8. The method for calculating air separation production using tank liquid level and its changes according to claim 1, characterized in that: The production data fusion algorithm module in step S4 also includes a system error compensation sub-step, which includes: S44. Periodically calibrate the differential pressure level gauge using a standard flow meter and record the output calculated by the differential pressure level gauge. Production measured by a standard flow meter Calculate the systematic error: ； in, This is the systematic error of the differential pressure level gauge. The output calculated by the differential pressure level gauge. Production volume measured by a standard flow meter; S45. Substitute the systematic error into the total output formula to correct the error: ； in, This represents the final total output after error compensation.

9. The method for calculating air separation production using tank liquid level and its changes according to claim 1, characterized in that: The steps in step S1 for collecting historical liquid level data during intermittent filling of the storage tank include: S11. Collect historical liquid level data for a specified duration, where "specified duration" is the time span of historical data that meets the algorithm parameter fitting requirements; S12. Historical data must include the number of fillings, filling start time, filling end time, real-time liquid level during filling, and ambient temperature for the corresponding time period. Among them, "number of fillings" refers to the total number of tank fillings per unit time, "real-time liquid level" refers to the liquid level data collected at fixed intervals during filling, and "ambient temperature" refers to the ambient temperature during the filling time period. S13. Clean the historical data, removing invalid and missing data. "Invalid data" refers to liquid level values ​​that are outside the normal range of the storage tank, and "missing data" refers to liquid level data that was not collected due to equipment failure.

10. The method for calculating air separation production using tank liquid level and its changes according to claim 1, characterized in that: Step S4, which integrates the algorithm module into the mainboard of the intelligent differential pressure level gauge, includes: S46. Compile the logic code of the dynamic liquid level production mapping algorithm module and the multi-component time series decomposition algorithm module into machine code executable by the microcontroller unit of the intelligent differential pressure liquid level gauge motherboard. Here, "logic code" is the program code that implements the algorithm function, and "microcontroller unit" is the core control component of the intelligent differential pressure liquid level gauge motherboard. S47. The compiled machine code is burned into the storage component of the microcontroller unit through the debugging interface, where "debugging interface" is a hardware interface used for code burning and program debugging, and "storage component" is a component in the microcontroller unit used to store program code; S48. After the programming is completed, simulate two scenarios: no liquid consumption and intermittent filling. Input test liquid level data to verify the production calculation results output by the motherboard. Among them, "test liquid level data" is the liquid level value generated by simulating the actual working conditions, and "production calculation results" is the production data output by the motherboard according to the algorithm module.

Images