Analyzer application method based on process requirements of air separation device

Abstract

The invention discloses an analysis meter application method based on process requirements of an air separation device, and relates to the technical field of analysis meter application, pressure fluctuation data and accuracy data of different gas sampling are obtained, and the installation positions of different types of analysis meters are determined through fuzzy reasoning; establishing a correlation model for the analyzer with the determined installation position; when an adjacent analyzer triggers a wake-up demand, acoustic characteristics in the air separation device are obtained by using the acoustic sensor array, a prediction model is constructed by combining the acoustic characteristics and the correlation model, and whether the current analyzer needs to be awakened or not is judged through the prediction model; the method comprises the following steps: identifying a change trend of gas characteristics collected by the awakened analyzer by using a time sequence, judging whether sampling parameters of the awakened analyzer need to be adjusted or not according to the change trend, and further outputting the sampling parameters of the awakened analyzer based on a preset parameter adjustment mechanism. Accurate positioning, intelligent linkage and dynamic adjustment based on data driving are achieved, and the detection efficiency and operation stability of the air separation device are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of analyzer application, and in particular to an analyzer application method based on process requirements of an air separation unit. Background Art

[0002] Air separation units (ASUs) are industrial equipment used to separate various components of air to produce industrial gases such as oxygen and nitrogen. They are widely used in industries such as metallurgy, petrochemicals, coal chemical industry, glass, semiconductors, aerospace, and nuclear power. Due to the deteriorating overall environment, the number of hazardous components in the air that threaten the safety of ASUs is increasing, such as hydrocarbons and nitrogen oxides. Acetylene and nitrous oxide are currently recognized as the most dangerous. These harmful components, when entering an ASU and accumulating over time and reaching a critical concentration, can cause an explosion within the unit, resulting in significant loss of life and property. To ensure safe operation, ASUs require the installation of various online analyzers, such as carbon dioxide and total hydrocarbon analyzers. These analyze trace levels of hydrocarbons, carbon dioxide, and nitrous oxide in the ASU, monitor the changes in the levels of these explosive components, and provide crucial data for the unit's safe operation.

[0003] Existing technologies select analyzer locations based on experience, without combining pressure fluctuations with sampling accuracy for quantitative analysis. Furthermore, the data from adjacent analyzers lacks effective correlation, making it impossible to predict detection needs in advance, resulting in wasted resources or delayed detection. The sampling frequency and measurement range are fixed and cannot be dynamically adjusted as gas characteristics change, affecting detection accuracy. Furthermore, there is a lack of non-invasive monitoring methods based on acoustic characteristics.

[0004] Therefore, in response to the above problems, there is an urgent need for an analyzer application method based on the process requirements of the air separation unit. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the present invention provides an analyzer application method based on the process requirements of the air separation unit, which solves the problems of blind installation, poor coordination, rigid parameters and insufficient adaptability to working conditions of the existing air separation unit analyzers.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an analyzer application method based on the process requirements of an air separation unit, comprising the following steps: step S1, obtaining pressure fluctuation data of different parts of the air separation unit during the sampling process, as well as accuracy data for different gas sampling, and obtaining installation scores of different types of analyzers in different parts of the air separation unit through fuzzy reasoning output, and determining the installation locations of different types of analyzers; step S2, for the analyzer with a determined installation location, identifying the correlation between the gas characteristics to be collected and the gas characteristics collected by the adjacent analyzer, and establishing a correlation model; step S3, when the adjacent analyzer triggers a wake-up demand, using an acoustic sensor array to obtain the acoustic characteristics in the air separation unit, combining the acoustic characteristics with the correlation model to construct a prediction model, and judging whether the current analyzer needs to be awakened by the prediction model; step S4, using a time series to identify the changing trend of the gas characteristics collected by the awakened analyzer, judging whether the sampling parameters of the awakened analyzer need to be adjusted based on the changing trend, and then outputting the sampling parameters of the awakened analyzer based on a preset parameter adjustment mechanism.

[0007] Furthermore, step S1 is specifically analyzed as follows: obtaining pressure fluctuation data of different parts of the air separation unit during the sampling process through a pressure sensor array, and obtaining accuracy data of different gas sampling in combination with a process requirement database, and then dividing the pressure fluctuation data and the accuracy data of different gas sampling into different fuzzy sets, and establishing fuzzy rules. The fuzzy rules take the pressure fluctuation data and the accuracy data of different gas sampling as input, and output a fuzzy set of installation scores of different types of analyzers in various parts. The fuzzy set of installation scores of different types of analyzers in various parts is defuzzified to obtain installation scores of different types of analyzers in various parts, and the parts of the air separation unit corresponding to the highest installation scores are respectively recorded as the installation locations of different types of analyzers.

[0008] Furthermore, the association model is specifically established and analyzed as follows: Granger causality is used to determine the influence direction and intensity between the gas characteristics collected by each analyzer, a preliminary association network is established, and a conditional trigger rule library is established to dynamically adjust the parameter weights of the association model according to the air separation unit load rate and product purity.

[0009] Furthermore, the specific analysis of constructing a prediction model by combining acoustic features with an association model is as follows: extracting training samples based on historical data, wherein the characteristics of the training samples include trigger parameters and acoustic features based on an association model; using random forest as the basis for constructing a prediction model, setting random forest parameters, optimizing the prediction model performance through cross-validation, and probabilistically calibrating the wake-up probability output by the prediction model so that the error between the wake-up probability and the actual wake-up situation is lower than an error threshold; deploying the prediction model, outputting the wake-up probability of the current analyzer, and waking up the current analyzer when the wake-up probability exceeds the wake-up probability threshold.

[0010] Furthermore, the trigger parameter based on the association model is specifically represented by: the association model outputs the influence intensity between the gas characteristics collected by the adjacent analyzer and the current analyzer, and the influence intensity threshold is stored in the conditional trigger rule library. When the influence intensity between the gas characteristics collected by the adjacent analyzer and the current analyzer exceeds the influence intensity threshold, the difference between the influence intensity and the influence intensity threshold is obtained as the trigger parameter of the association model.

[0011] Furthermore, step S4 is specifically analyzed as follows: using time series to identify the changing trend of the gas characteristics collected by the awakened analyzer, judging whether the sampling parameters need to be adjusted by extreme point detection and slope in the sliding window, and if the sampling parameters need to be adjusted, using the parameter adjustment mechanism to output the sampling parameters of the awakened analyzer, the sampling parameters including the sampling frequency and the sampling range; the parameter adjustment mechanism is specifically: based on the historical data and the air separation unit process requirements, the changing trend of the collected gas characteristics is identified and the influence association between the sampling frequency adjustment ratio and the sampling range adjustment ratio.

[0012] Furthermore, the specific analysis of determining whether the sampling parameters need to be adjusted through extreme point detection and slope within the sliding window is as follows: setting the sliding window size, performing first-order difference processing on the gas characteristics within the sliding window, and obtaining a change rate sequence; detecting the local maximum and minimum values ​​of the change rate sequence, and if the number of extreme points is greater than or equal to 2 and the difference between the local maximum and minimum values ​​exceeds a preset difference threshold, determining that the change trend of the collected gas characteristics is a significant change trend; and using slope analysis to output the change trend type, the change trend type includes rising, falling, and fluctuating.

[0013] Furthermore, it also includes parameter optimization verification using the digital twin model. The specific steps include: building a digital twin model of the air separation unit, integrating real-time analyzer data and equipment status data, the real-time analyzer data includes collected gas characteristics and sampling parameters, and the equipment status data includes equipment temperature and equipment pressure; simulating the parameter adjustment effect in the digital twin, comparing the deviation between the actual air separation unit and the digital twin model, if the deviation exceeds the deviation threshold, triggering parameter correction, the parameter correction includes adjusting the gas diffusion coefficient and equipment response delay; verifying the corrected parameter adjustment effect through the digital twin model, and sending the critical value of the corresponding parameter when the deviation between the actual air separation unit and the digital twin model is lower than the deviation threshold as the parameter adjustment plan to the actual analyzer for execution, forming a parameter closed-loop optimization.

[0014] The present invention has the following beneficial effects: This analyzer application method based on the process requirements of the air separation unit and the multi-dimensional fusion installation location optimization method break through the limitations of traditional single indicator selection, and realizes scientific decision-making on the installation location through the combination of fuzzy reasoning and integer programming; dynamic correlation modeling technology solves the problem of poor adaptability of fixed correlation models, and improves the collaborative efficiency of the analyzer through process state identification and abnormal propagation prediction; the acoustic feature-assisted wake-up mechanism creatively converts equipment operation noise into a useful monitoring signal, which has higher sensitivity and reliability compared with traditional timed wake-up or single parameter triggering; the adaptive parameter adjustment method realizes dynamic optimization of sampling parameters through reinforcement learning, reducing system energy consumption while ensuring measurement accuracy; digital twin assisted decision-making deeply integrates virtual simulation with actual monitoring, realizes early detection and accurate prediction of faults, and provides technical support for preventive maintenance; self-optimization capability enables it to continuously adapt to process changes and equipment aging, extends the effective service cycle of the system, and reduces operation and maintenance costs.

[0015] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The present invention is a flow chart of the analyzer application method based on the process requirements of the air separation unit. DETAILED DESCRIPTION

[0017] The embodiment of the present application realizes data-driven precise positioning, intelligent linkage, and dynamic adjustment through an analyzer application method based on the process requirements of the air separation unit, thereby improving the detection efficiency and operational stability of the air separation unit.

[0018] The overall idea of ​​the embodiment of the present application is: quantify the installation position of the analyzer based on fuzzy reasoning; use the association model to mine the value of proximity detection data; realize intelligent wake-up of the analyzer through acoustic features and prediction models; dynamically adjust the sampling parameters in combination with time series to form a "positioning-linkage-wake-up-adjustment" closed loop.

[0019] See also Figure 1, an embodiment of the present invention provides a technical solution: an analyzer application method based on the process requirements of an air separation unit, comprising the following steps: step S1, obtaining pressure fluctuation data of different parts of the air separation unit during the sampling process, as well as accuracy data for different gas sampling, and obtaining installation scores of different types of analyzers in different parts of the air separation unit through fuzzy reasoning output, and determining the installation locations of different types of analyzers; step S2, for the analyzer with the determined installation location, identifying the correlation between the gas characteristics to be collected by the analyzer and the gas characteristics collected by the adjacent analyzer, and establishing a correlation model; step S3, when the adjacent analyzer triggers a wake-up demand, using an acoustic sensor array to obtain the acoustic characteristics within the air separation unit, combining the acoustic characteristics with the correlation model to construct a prediction model, and judging whether the current analyzer needs to be awakened by the prediction model; step S4, using a time series to identify the change trend of the gas characteristics collected by the awakened analyzer, judging whether the sampling parameters of the awakened analyzer need to be adjusted based on the change trend, and then outputting the sampling parameters of the awakened analyzer based on a preset parameter adjustment mechanism.

[0020] Specifically, step S1 is specifically analyzed as follows: obtaining pressure fluctuation data of different parts of the air separation unit during the sampling process through a pressure sensor array, and obtaining accuracy data of different gas sampling in combination with a process requirement database, and then dividing the pressure fluctuation data and the accuracy data of different gas sampling into different fuzzy sets, and establishing fuzzy rules. The fuzzy rules take the pressure fluctuation data and the accuracy data of different gas sampling as input, and output a fuzzy set of installation scores of different types of analyzers in various parts. The fuzzy set of installation scores of different types of analyzers in various parts is defuzzified to obtain installation scores of different types of analyzers in various parts, and the parts of the air separation unit corresponding to the highest installation scores are recorded as the installation locations of different types of analyzers.

[0021] In this implementation plan, pressure fluctuation data includes pressure fluctuation amplitude, pressure change frequency, and pressure gradient distribution. The pressure fluctuation amplitude represents the maximum deviation of the pressure value from the average pressure per unit time, reflecting gas flow stability (e.g., a sudden pressure increase in a distillation tower may cause compositional disturbances); the pressure change frequency represents the number of pressure fluctuations per unit time, reflecting the dynamic characteristics of the process (e.g., the frequency of pressure changes is higher during compressor startup and shutdown); and the pressure gradient distribution represents the pressure difference between different parts of the air separation unit, used to determine gas flow direction and diffusion characteristics (e.g., the pressure gradient at the inlet and outlet of a heat exchanger affects sampling representativeness). Pressure fluctuation data is acquired by deploying a pressure sensor array at key locations in the air separation unit (e.g., compressor outlet, distillation tower trays, and heat exchanger inlet and outlet). Analog signal conditioning circuits (amplification and filtering) acquire pressure signals in real time at a sampling frequency of no less than 10 Hz. These analog pressure signals are converted to digital signals using an analog-to-digital converter (ADC) and transmitted to a data processing module via industrial Ethernet or a fieldbus (e.g., PROFINET or Modbus). A Kalman filter algorithm is used to denoise the raw pressure data to eliminate the effects of noise such as pipeline vibration and electromagnetic interference on the calculation of fluctuation parameters.

[0022] The accuracy data of different gas sampling include target gas concentration deviation, component response time, and cross-interference coefficient, among which: the target gas concentration deviation represents the absolute value of the difference between the actual sampling concentration and the concentration required by the process, reflecting the sampling ability to capture the target gas (such as the impurity content detection deviation in high-purity oxygen sampling); the component response time represents the reaction speed of the analyzer to changes in gas components, reflecting the efficiency of the sampling pipeline and pretreatment (such as the control lag caused by the long lag time in trace water detection); the cross-interference coefficient represents the degree of interference of non-target gases on the test results, which is used to evaluate the targeted sampling (such as the interference of nitrogen mixed in argon on the thermal conductivity analyzer). The accuracy data of different gas sampling is obtained as follows: the theoretical concentration range, allowable deviation threshold and interfering gas type of the target gas in each part are extracted from the process design documents of the air separation unit, and a standardized parameter table is established. The detection data of the analyzer under typical operating conditions can also be collected, and the target gas concentration deviation and component response time are calculated. The sampling accuracy characteristic curve of each part is obtained by fitting the least squares method, or the operating conditions of the air separation unit are simulated in a laboratory environment, and interfering gases of different proportions are injected to determine the cross-interference coefficient to form an interference matrix and store it in the process requirements database.

[0023] The specific analysis of the output of the fuzzy set of installation scores of different types of analyzers in various parts is as follows: the pressure fluctuation data and the accuracy data of different gas sampling are defined as input variables, and divided into different fuzzy sets, for example: "Low", "Medium", "High" for pressure fluctuation data, "Low", "Medium", "High" for different gas sampling accuracy data, the installation scores of different types of analyzers in various parts are defined as output variables, and divided into different fuzzy sets, for example: "Low", "Medium", "High" for different types of analyzers in various parts. Fuzzy rules are established based on the fact that the accuracy data of different gas sampling is proportional to the installation scores of different types of analyzers in various parts, and the pressure fluctuation data is inversely proportional to the installation scores of different types of analyzers in various parts. Examples of fuzzy rules are: Marking the accuracy data of different gas sampling as Q, the pressure fluctuation data as G, and the installation scores of different types of analyzers in various locations as R, we can define: Rule 1: IF (Q is High) AND (G is Short) THEN (R is High) Rule 2: IF (Q is Short) AND (G is High) THEN (R is Low) ... It should be noted that the division of fuzzy sets can be adjusted according to actual conditions. For example, although this embodiment takes three fuzzy sets as an example, in fact, the pressure fluctuation data, the accuracy data of different gas sampling, and the installation scores of different types of analyzers in various locations can be divided into more than three sets to facilitate better installation effects of different types of analyzers in various locations.

[0024] For the judgment of high, medium and low pressure fluctuation data and high, medium and low accuracy data of different gas sampling, thresholds can be set according to actual conditions for judgment, which will not be elaborated here.

[0025] Through fuzzy reasoning, pressure fluctuations and sampling accuracy are quantified into installation scores, avoiding the subjectivity of traditional empirical methods, ensuring that the analyzer installation location matches process requirements, and reducing detection errors caused by improper positioning; combining the two-dimensional data of pressure fluctuations and sampling accuracy to cover the fluid characteristics and detection requirements of different parts of the air separation unit, and enhance the adaptability of the analysis system to complex working conditions such as high pressure, low pressure, and impurity fluctuations; through clear fuzzy set partitioning and defuzzification steps, a reusable algorithm framework is formed, which facilitates the rapid deployment of analyzer networks in different air separation units and improves system integration efficiency.

[0026] Specifically, the association model is established and analyzed as follows: Granger causality is used to determine the direction and intensity of influence between the gas characteristics collected by each analyzer, a preliminary association network is established, and a conditional trigger rule library is established to dynamically adjust the parameter weights of the association model according to the air separation unit load rate and product purity.

[0027] In this implementation scheme, the specific steps of Granger causality analysis are as follows: denoising and normalizing the gas characteristic time series collected by each analyzer to eliminate the influence of dimensional differences and random noise; using the ADF test or KPSS test to determine whether the time series is stationary, and performing difference processing on non-stationary series until it is stationary; using information criteria (such as AIC and BIC) to determine the optimal lag order of the Granger causality test to ensure a balance between model complexity and fitting effect; performing a two-way Granger causality test on each pair of gas characteristic time series to determine whether there is a unidirectional or bidirectional causal relationship; determining the intensity level of the causal relationship based on the size of the F statistic value, and generating an association network topology structure that includes the direction and intensity of influence; verifying the significance of the causal relationship through Monte Carlo simulation or the Bootstrap method to eliminate pseudo causal relationships.

[0028] The specific steps for constructing a conditional trigger rule library are as follows: based on historical operating data and process safety standards, set initial impact intensity thresholds for each association; classify and store thresholds according to process stages (such as startup, steady state, and shutdown) and operating conditions (such as high load and low purity); assign priorities to each trigger rule to ensure that they can be executed in order of importance when multiple rules are met at the same time; regularly collect actual operating data, optimize threshold parameters through machine learning algorithms, and improve the accuracy of trigger rules; set a fault-tolerant mechanism, and when the trigger parameters exceed the preset range, automatically switch to safe mode and record abnormal logs; establish a graphical interface for the rule library to support manual intervention and rule editing, and realize human-machine collaborative optimization.

[0029] The specific steps for dynamically adjusting parameter weights are as follows: real-time data collection of air separation unit load rate and product purity is used to determine the current operating condition through a pattern recognition algorithm; a predefined operating condition-weight mapping table is used to match the corresponding parameter weight combination based on the current operating condition; when switching between operating conditions, a weighted average method is used to achieve a smooth transition of parameter weights to avoid drastic model fluctuations; the prediction results of the correlation model are compared with the actual test data, and the weight parameters are dynamically adjusted based on the error; a manual intervention interface can also be set up to allow process experts to manually adjust the weight parameters based on their experience. An example of a conditional trigger rule library is: when a nearby analyzer detects that the argon content exceeds a preset argon content threshold, the oxygen analyzer is triggered to enter a wake-up state.

[0030] Granger causal analysis is used to determine the causal relationship between gas characteristics, avoiding the pseudo-correlation problem caused by traditional methods that rely solely on correlation analysis, thereby improving the reliability of model predictions. A conditional trigger rule library is established to achieve intelligent wake-up of the analyzer, reduce detection, and reduce system energy consumption and equipment loss. Parameter weights are dynamically adjusted according to load rate and product purity, so that the correlation model can adapt to different operating conditions of the air separation unit and improve system robustness.

[0031] Specifically, the specific analysis of building a prediction model by combining acoustic features with the association model is as follows: extracting training samples based on historical data, and the characteristics of the training samples include trigger parameters and acoustic features based on the association model; using random forest as the basis for building a prediction model, setting random forest parameters, and optimizing the prediction model performance through cross-validation, probabilistically calibrating the wake-up probability output by the prediction model so that the error between the wake-up probability and the actual wake-up situation is lower than the error threshold; deploying the prediction model, outputting the wake-up probability of the current analyzer, and waking up the current analyzer when the wake-up probability exceeds the wake-up probability threshold.

[0032] In this implementation scheme, the trigger parameters based on the association model are specifically represented; the association model outputs the influence intensity between the gas characteristics collected by the adjacent analyzer and the current analyzer, and the influence intensity threshold is stored in the conditional trigger rule library. When the influence intensity between the gas characteristics collected by the adjacent analyzer and the current analyzer exceeds the influence intensity threshold, the difference between the influence intensity and the influence intensity threshold is obtained as the trigger parameter of the association model.

[0033] Acoustic characteristics include frequency distribution characteristics, time domain characteristics, spatial characteristics, and statistical characteristics. Among them, frequency distribution characteristics include main frequency components, harmonic content, frequency band energy distribution, etc., which reflect the gas flow state and equipment operation state; time domain characteristics represent sound pressure level, sound intensity, sound signal duration, rising edge / falling edge time, etc., which reflect the intensity and change rate of acoustic events; spatial characteristics represent the phase difference and sound pressure gradient between channels of the acoustic sensor array, which are used to locate the sound source and determine the sound wave propagation path; statistical characteristics represent the mean, variance, kurtosis, skewness, etc. of the acoustic signal, which describe the statistical characteristics and non-stationarity of the signal. The steps for acquiring acoustic features are as follows: arrange no less than four high-sensitivity acoustic sensors in a ring at key locations of the air separation unit (such as pipe elbows, valves, and tower connections) to form a spatial array; synchronously collect acoustic signals from each sensor at a sampling frequency of no less than 20 kHz, with a duration of no less than 10 seconds per time; perform bandpass filtering on the original acoustic signal to remove interference from environmental noise and the inherent frequency of the equipment; use time-frequency analysis methods to extract the time domain, frequency domain, and spatial feature parameters of the acoustic signal; screen the subset of acoustic features that contribute most to the wake-up decision through correlation analysis or feature importance ranking; and normalize the extracted acoustic features to eliminate dimensionality effects.

[0034] Acoustic characteristics can reflect the fluid dynamics characteristics (such as airflow disturbance and vortex formation) and mechanical conditions (such as valve vibration and pipeline friction) within the air separation unit. These microscopic changes often appear earlier than changes in gas composition and can be used as early warning indicators to achieve forward-looking decisions on analyzer wake-up.

[0035] The steps for predicting the model to output the wake-up probability are as follows: extract the association model trigger parameters and the acoustic features at the corresponding moment from historical data, mark whether the actual wake-up occurs, and form a training data set; use the random forest algorithm to build a prediction model, and set initial parameters such as the number of decision trees, maximum depth, and minimum number of sample splits; divide the training data into multiple subsets, use the K-fold cross-validation method to evaluate the model performance, and adjust the parameters to optimize indicators such as accuracy and recall; use Platt scaling or Isotonic regression to calibrate the original probability output by the model to ensure that the calibrated probability is consistent with the actual wake-up frequency; use an independent test data set to verify the error rate of the calibrated model to ensure that the error is lower than the preset threshold; input the association model trigger parameters and acoustic features at the current moment into the calibrated prediction model to output the wake-up probability.

[0036] The wake-up probability threshold is the critical value for determining whether to wake up the analyzer. When the value is higher than the threshold, it is considered that the analyzer needs to be woken up for detection. When the value is lower than the threshold, the analyzer is maintained in standby mode to save resources. The method for obtaining the value is: analyze the false alarm rate and missed alarm rate under different wake-up probabilities in the historical operation data, and select the probability value that minimizes the sum of the two as the initial threshold; the energy consumption cost of waking up the analyzer, the loss caused by detection delay, and the maintenance cost caused by false alarms can also be comprehensively considered to establish a cost function to solve the optimal threshold; the threshold can also be dynamically adjusted through a preset mapping relationship based on operating parameters such as the air separation unit load rate and product purity requirements; process experts are allowed to manually correct the threshold based on experience, forming a threshold optimization mechanism for human-machine collaboration.

[0037] The impact intensity threshold is obtained in the following ways: based on the design parameters and operating manual of the air separation unit, the theoretical correlation intensity threshold between each gas characteristic is determined; or the correlation intensity distribution under normal and abnormal conditions in the historical operating data is analyzed, and the upper limit of the 95% confidence interval of the normal condition is taken as the initial threshold; the system performance changes can also be observed by changing the threshold parameters, and the threshold point with the highest sensitivity to the awakening decision can be selected; the online learning algorithm can also be used to continuously optimize the impact intensity threshold according to the latest operating data to adapt it to changes such as device aging and process adjustments.

[0038] By combining acoustic features with associated model trigger parameters, the limitations of single gas detection are overcome and the ability to perceive potential changes in air separation units at an early stage is enhanced. By optimizing the prediction model through cross-validation and probability calibration, the reliability of the wake-up decision is ensured and the false alarm and missed alarm rates are reduced. The wake-up threshold is dynamically adjusted based on real-time acoustic features, so that the analyzer can maintain the optimal response state under different working conditions and reduce energy waste.

[0039] Specifically, step S4 is analyzed as follows: using time series to identify the changing trend of the gas characteristics collected by the awakened analyzer, judging whether the sampling parameters need to be adjusted through extreme point detection and slope in the sliding window, and if the sampling parameters need to be adjusted, using the parameter adjustment mechanism to output the sampling parameters of the awakened analyzer, the sampling parameters include sampling frequency and sampling range; the parameter adjustment mechanism is specifically: based on historical data and the air separation unit process requirements, the changing trend of the collected gas characteristics is identified and the influence association between the sampling frequency adjustment ratio and the sampling range adjustment ratio.

[0040] In this embodiment, the specific analysis of whether the sampling parameters need to be adjusted by detecting extreme points and slope in the sliding window is as follows: setting the sliding window size, performing first-order difference processing on the gas characteristics in the sliding window, and obtaining a change rate sequence; detecting the local maximum and minimum values ​​of the change rate sequence, if the number of extreme points is greater than or equal to 2 and the difference between the local maximum and minimum values ​​exceeds the preset difference threshold, it is determined that the change trend of the collected gas characteristics is a significant change trend; using slope analysis to output the change trend type, the change trend types include rising, falling, and fluctuating.

[0041] The steps of using time series to identify change trends are as follows: receiving gas characteristic time series data collected by the awakened analyzer, cleaning the data, removing outliers and missing values, and unifying the data dimension using a normalization method; setting an appropriate sliding window size based on the process characteristics of the air separation unit and the fluctuation of historical data, and dividing the time series data into multiple overlapping or non-overlapping window data segments; performing first-order difference calculation on the gas characteristic data in each sliding window to obtain a change rate sequence reflecting the data change rate; traversing the change rate sequence, detecting local maximum and minimum points, and recording the positions and values ​​of the extreme points; counting the number of detected extreme points, and calculating the difference between adjacent extreme points. When the number of extreme points is greater than or equal to 2 and the difference between the local maximum and minimum exceeds a preset difference threshold, the change trend of the gas characteristics collected in the window is determined to be a significant change trend; otherwise, the change trend is deemed to be insignificant; the judgment results of all sliding windows are summarized and analyzed. If multiple windows are determined to be significant change trends, the current gas characteristics are determined to have shown significant changes. Otherwise, the change trend is determined to be stable.

[0042] The sampling frequency refers to the number of times the analyzer collects gas characteristics per unit time. When the gas characteristics change slowly, lowering the sampling frequency can reduce the equipment operating load and data processing volume, saving energy. When the gas characteristics change dramatically, such as during the startup and load adjustment stages of the air separation unit, increasing the sampling frequency can capture data changes more intensively, ensuring that key information is not missed, and providing timely and accurate data support for process control.

[0043] The sampling range indicates the numerical range of gas characteristic parameters that the analyzer can detect. By adjusting the sampling range, the detection range can be reasonably set according to the actual changes in gas characteristics. When the gas characteristic values ​​are close to or exceed the original sampling range, expanding the sampling range can avoid data overflow or inaccurate measurement. When the gas characteristic values ​​are relatively stable and concentrated in a smaller range, narrowing the sampling range can improve detection accuracy and make the measurement results more accurate.

[0044] The steps for constructing the parameter adjustment mechanism are as follows: 1. Collect the gas characteristic change trend data collected by the awakened analyzer under different operating conditions of the air separation unit, as well as the adjustment records of the sampling frequency and sampling range under the corresponding operating conditions, and establish a historical database; 2. Combine the design documents, operating procedures and process expert experience of the air separation unit to clarify the accuracy, frequency and other requirements of gas characteristic detection in different process stages and operating conditions, and compile a process requirement parameter table; 3. Use data mining algorithms to analyze the potential correlation between the gas characteristic change trend (increase, decrease, fluctuation) in the historical data and the sampling frequency adjustment ratio and sampling range adjustment ratio, and determine the adjustment rules and amplitude range of the sampling parameters under different change trends; 4. Use historical data to simulate and verify the association relationship and adjustment rules obtained by mining, and evaluate the rationality of the rules by comparing the accuracy and effectiveness of the detection data before and after adjustment; 5. Based on the verification results, optimize the association relationship and rules until they meet the process requirements and detection accuracy requirements; 6. Integrate the optimized association relationship and adjustment rules into the parameter adjustment mechanism so that it can automatically calculate and output the corresponding sampling frequency adjustment ratio and sampling range adjustment ratio based on the real-time identification of the gas characteristic change trend.

[0045] The sliding window size is set as follows: Based on the process characteristics of the air separation unit and previous operating experience, process experts or technicians set an initial sliding window size as a reference value for the initial system operation. During system operation, the fluctuation and frequency of gas characteristic data are monitored in real time. When data fluctuations are large and the frequency of change is high, the sliding window size is appropriately reduced to more accurately capture data changes. When data fluctuations are small and the frequency of change is low, the sliding window size is increased to reduce the computational workload and data processing pressure. Historical data can also be regularly analyzed to calculate the accuracy and effectiveness of time series analysis (such as detection error rate and trend identification accuracy) under different sliding window sizes. Based on the analysis results, optimization algorithms (such as genetic algorithms and particle swarm optimization) are used to search for the optimal sliding window size and continuously adjust and optimize the window size.

[0046] The steps for using slope analysis to output the change trend type are as follows: perform linear fitting on the gas characteristic data in the sliding window and calculate the slope of the fitted line, which reflects the average change trend of the gas characteristic data in the window; make a judgment based on the calculated slope value. When the slope is greater than 0, the change trend type is determined to be rising; when the slope is less than 0, the change trend type is determined to be falling; when the slope is close to 0 and fluctuates within a certain error range, combined with the extreme point detection results, if there are multiple extreme points and the difference is small, the change trend type is determined to be fluctuating; set the slope threshold range. When the slope exceeds the threshold range of the normal rising or falling trend, it is regarded as an abnormal situation. Further combined with other data analysis methods (such as data mutation detection and historical data comparison) for comprehensive judgment to ensure the accuracy of the change trend type judgment.

[0047] Through time series analysis and sliding window detection, it is possible to keenly capture subtle changes in gas characteristics, determine the need for sampling parameter adjustment in real time, ensure that the analyzer detection data accurately reflects the operating status of the air separation unit, and avoid detection lag or data distortion caused by fixed parameters; dynamically adjust the sampling frequency and range according to the changing trend of gas characteristics, reduce the sampling frequency when the gas characteristics are stable, and reduce equipment operating losses and data redundancy; increase the sampling frequency and expand the sampling range when the changes are significant to ensure the complete collection of key data and achieve efficient use of resources; the parameter adjustment mechanism combines historical data with process requirements to establish an impact association, so that the system can adapt to different operating conditions and process fluctuations of the air separation unit, improve the stability and reliability of the analyzer in complex environments, and ensure the safe and efficient operation of the air separation unit.

[0048] Specifically, it also includes parameter optimization verification using the digital twin model. The specific steps include: building a digital twin model of the air separation unit, integrating real-time analyzer data and equipment status data, where the real-time analyzer data includes collected gas characteristics and sampling parameters, and the equipment status data includes equipment temperature and equipment pressure; simulating the parameter adjustment effect in the digital twin, and comparing the deviation between the actual air separation unit and the digital twin model. If the deviation exceeds the deviation threshold, parameter correction is triggered, which includes adjusting the gas diffusion coefficient and equipment response delay; verifying the corrected parameter adjustment effect through the digital twin model, and sending the critical value of the corresponding parameter when the deviation between the actual air separation unit and the digital twin model is lower than the deviation threshold as the parameter adjustment plan to the actual analyzer for execution, forming a closed-loop parameter optimization.

[0049] In this implementation plan, the steps for constructing a digital twin model of an air separation unit are as follows: based on the design drawings and 3D scanning data of the air separation unit, a 3D geometric model of the unit is constructed using professional modeling software, covering core equipment and connection structures such as air compressors, distillation towers, heat exchangers, and pipelines, to ensure that the model is consistent with the actual unit's geometric dimensions; each component in the model is assigned material properties (such as density and thermal conductivity), thermodynamic parameters (specific heat capacity, latent heat of phase change), and fluid mechanics parameters (viscosity and diffusion coefficient), with parameter values ​​referenced by design documents and experimental test data; the automated control logic of the air separation unit is sorted out, and control algorithms and strategies such as temperature control, pressure regulation, and flow control are transplanted into the digital twin model to establish a virtual control module corresponding to the actual control system; a standardized data interface is designed to enable data interaction with real-time analyzers, equipment sensors, and control systems to ensure that the model can receive real-time detection data and control instructions; after the model is built, historical operating data is input, and the model output results are compared with the actual unit's historical data. By adjusting model parameters and correcting logical relationships, the model output error is brought within an acceptable range, completing model calibration.

[0050] Real-time analyzer data directly reflects the gas composition and detection status in the air separation unit, providing real-time process parameters for the digital twin model, enabling the model to simulate based on actual operating conditions, and ensuring that the simulation results are closely related to the actual operating status; real-time updated analyzer data is used to monitor process change trends. When gas characteristics fluctuate abnormally, the model can capture the changes in time and simulate the effects of different parameter adjustment schemes, providing data support for parameter optimization and realizing dynamic optimization of the operation of the air separation unit.

[0051] Equipment status data reflects the operating status of the equipment. Integrating this data enables the digital twin model to monitor the health status of the equipment in real time, detect potential faults such as equipment overheating and abnormal pressure in advance, provide early warning information for equipment maintenance, and avoid unplanned downtime; equipment status directly affects the process performance of the air separation unit. For example, changes in heat exchanger temperature will affect the gas heat exchange efficiency, and compressor pressure fluctuations will change the gas flow rate. Integrating equipment status data into the model enables the model to more realistically simulate the interaction between equipment and process, thereby improving the accuracy of simulation results.

[0052] The deviation threshold is a critical indicator that measures the difference between the actual ASU and the digital twin's operating state. It's used to determine the degree of consistency between the model simulation results and the actual situation. When the deviation exceeds this threshold, it indicates poor model simulation or abnormal operation of the actual unit, triggering parameter correction or operational adjustment. Domain experts can set initial deviation thresholds based on ASU design standards, process requirements, and historical operating experience. For example, the gas concentration deviation threshold can be set to ±2% of the actual value, and the equipment pressure deviation threshold to ±5%. Alternatively, actual data from the ASU under stable operating conditions and simulated model data can be collected, and the deviations between the two in various parameter dimensions (gas composition, equipment pressure, temperature, etc.) can be calculated. Statistical analysis (such as calculating standard deviations and confidence intervals) can be used to determine an appropriate deviation threshold range. A dynamic deviation threshold adjustment mechanism can also be established to monitor the deviation between the actual unit and the model and the effectiveness of adjustments in real time. If the deviation frequently exceeds the threshold after multiple adjustments, the threshold range can be automatically narrowed. If the system is stable and the deviation is small after adjustments, the threshold range can be appropriately expanded to balance parameter adjustment sensitivity with system stability.

[0053] Adjusting the gas diffusion coefficient and device response delay is crucial for the following reasons: The gas diffusion coefficient affects the mixing and separation processes within an air separation unit. Adjusting this coefficient allows the digital twin model to more accurately simulate gas diffusion within pipelines and towers, thereby optimizing key process indicators such as distillation efficiency and product purity and improving the model's simulation accuracy. When the feed gas composition and operating conditions of an air separation unit change, the gas diffusion characteristics will also change. Adjusting the gas diffusion coefficient allows the model to quickly adapt to these changes, providing reliable simulation results for process parameter optimization and ensuring stable operation of the unit under different operating conditions. Regarding device response delay, after receiving control commands, actual equipment experiences a certain response delay due to factors such as mechanical inertia and signal transmission delay. Adjusting the device response delay parameter aligns the device behavior in the digital twin model with the actual equipment, avoiding control deviations caused by differences between the model and actual responses and improving the effectiveness of the control strategy. Simulating control effects under different device response delays can also be used to evaluate the rationality of existing control strategies, providing a basis for improving control algorithms and adjusting control parameters, reducing control overshoot and oscillation, and ensuring smooth and efficient operation of the air separation unit.

[0054] In summary, this application has at least the following effects: The analyzer installation location is determined through fuzzy reasoning, and dynamic wake-up and parameter adjustment are achieved by combining the correlation between acoustic characteristics and gas characteristics, thereby improving detection accuracy and response speed. The correlation model and prediction model are used to realize the data linkage of adjacent analyzers, reduce redundant detection, and optimize the operating efficiency of the air separation unit. The sampling parameters are automatically adjusted based on time series analysis to minimize sampling errors and adapt to the complex working conditions of the air separation process.

[0055] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0056] The present invention is described with reference to flowcharts of methods according to embodiments of the present invention. It should be understood that each combination of processes in the flowcharts can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts. Figure 1 A device that specifies functions in a process or multiple processes.

[0057] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A function specified in a process or multiple processes.

[0058] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 The steps of a specified function in a process or multiple processes.

[0059] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0060] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. The analyzer application method based on the process requirements of the air separation unit is characterized by: The following steps are involved: Step S1: Obtain pressure fluctuation data at different locations of the air separation unit during sampling, as well as accuracy data for different gas sampling. Fuzzy inference output is used to obtain installation scores for different types of analyzers at different locations of the air separation unit, and determine the installation locations of different types of analyzers. Step S2: for the analyzer at the determined installation location, identifying the correlation between the gas characteristics to be collected by the analyzer and the gas characteristics collected by the adjacent analyzers, and establishing a correlation model; Step S3: When a nearby analyzer triggers a wake-up request, the acoustic sensor array is used to obtain the acoustic characteristics in the air separation unit, and a prediction model is constructed by combining the acoustic characteristics with the correlation model. The prediction model is used to determine whether the current analyzer needs to be woken up; Step S4, using the time series to identify the changing trend of the gas characteristics collected by the awakened analyzer, judging whether the sampling parameters of the awakened analyzer need to be adjusted based on the changing trend, and then outputting the sampling parameters of the awakened analyzer based on the preset parameter adjustment mechanism.

2. The analyzer application method based on the process requirements of the air separation unit according to claim 1 is characterized in that: The specific analysis of step S1 is as follows: The pressure fluctuation data of different parts of the air separation unit during the sampling process are obtained through a pressure sensor array, and the accuracy data of different gas sampling are obtained in combination with the process requirement database. The pressure fluctuation data and the accuracy data of different gas sampling are then divided into different fuzzy sets, and fuzzy rules are established. The fuzzy rules take the pressure fluctuation data and the accuracy data of different gas sampling as input, and output a fuzzy set of installation scores of different types of analyzers in various parts. The fuzzy sets of installation scores of different types of analyzers in various parts are defuzzified to obtain the installation scores of different types of analyzers in various parts, and the parts of the air separation unit corresponding to the highest installation scores are recorded as the installation locations of different types of analyzers.

3. The analyzer application method based on the process requirements of the air separation unit according to claim 1 is characterized in that: The specific establishment and analysis of the association model is as follows: Granger causality is used to determine the influence direction and intensity between the gas characteristics collected by each analyzer, establish a preliminary association network, and build a conditional trigger rule library to dynamically adjust the parameter weights of the association model according to the air separation unit load rate and product purity.

4. The analyzer application method based on the process requirements of the air separation unit according to claim 1 is characterized in that: The specific analysis of combining acoustic features with the association model to build a prediction model is as follows: Extracting training samples based on historical data, where the training sample features include trigger parameters and acoustic features based on the association model; Based on the random forest as the basis for the prediction model, the random forest parameters are set, the prediction model performance is optimized through cross-validation, and the awakening probability output by the prediction model is calibrated to ensure that the error between the awakening probability and the actual awakening situation is lower than the error threshold; Deploy the prediction model and output the wakeup probability of the current analyzer. When the wakeup probability exceeds the wakeup probability threshold, wake up the current analyzer.

5. The analyzer application method based on the process requirements of the air separation unit according to claim 3 is characterized in that: The trigger parameters based on the association model are specifically represented; The association model outputs the influence intensity between the gas characteristics collected by the adjacent analyzer and the current analyzer. The influence intensity threshold is stored in the conditional trigger rule library. When the influence intensity between the gas characteristics collected by the adjacent analyzer and the current analyzer exceeds the influence intensity threshold, the difference between the influence intensity and the influence intensity threshold is obtained as the trigger parameter of the association model.

6. The analyzer application method based on the process requirements of the air separation unit according to claim 1 is characterized in that: The specific analysis of step S4 is as follows: Using time series to identify the changing trend of the gas characteristics collected by the awakened analyzer, and judging whether the sampling parameters need to be adjusted by detecting extreme points and slopes within the sliding window, if the sampling parameters need to be adjusted, the parameter adjustment mechanism is used to output the sampling parameters of the awakened analyzer, which include the sampling frequency and sampling range; The parameter adjustment mechanism is specifically: the influence association between the change trend of the collected gas characteristics identified based on historical data and the process requirements of the air separation unit and the adjustment ratio of the sampling frequency and the sampling range.

7. The method for applying an analyzer based on the process requirements of an air separation unit according to claim 6, characterized in that: The specific analysis of determining whether the sampling parameters need to be adjusted by detecting extreme points and slopes within the sliding window is as follows: Set the sliding window size, perform first-order difference processing on the gas characteristics within the sliding window, and obtain the rate of change sequence; Detect the local maximum and minimum values ​​of the change rate sequence. If the number of extreme points is greater than or equal to 2 and the difference between the local maximum and minimum exceeds the preset difference threshold, the change trend of the collected gas characteristics is determined to be a significant change trend; The slope analysis is used to output the change trend type, which includes rising, falling, and fluctuating.

8. The analyzer application method based on the process requirements of the air separation unit according to claim 1 is characterized in that: It also includes parameter optimization verification using the digital twin model. The specific steps include: Build a digital twin model of the air separation unit, integrating real-time analyzer data and equipment status data. The real-time analyzer data includes collected gas characteristics and sampling parameters, and the equipment status data includes equipment temperature and equipment pressure. Simulating the effect of parameter adjustments in the digital twin and comparing the deviation between the actual air separation unit and the digital twin model. If the deviation exceeds a deviation threshold, triggering parameter correction, such as adjusting the gas diffusion coefficient and device response delay; The modified parameter adjustment effect is verified through the digital twin model. When the deviation between the actual air separation unit and the digital twin model is lower than the deviation threshold, the critical value of the corresponding parameter is sent to the actual analyzer as the parameter adjustment plan for execution, forming a closed-loop parameter optimization.

Images