Method for process phase identification of ultrahigh pressure vessels based on non-invasive temperature monitoring

Abstract

The application discloses a kind of based on non-invasive temperature monitoring's superhigh pressure container process phase identification method, it is related to superhigh pressure container monitoring technical field.The upper end cover of the working cylinder body of the superhigh pressure container is installed with at least two temperature sensors in non-pressure area, and the lower end cover is also installed with at least two temperature sensors in non-pressure area;The method comprises the following steps: synchronously collecting the temperature time series data of each temperature sensor;Extracting multidimensional temperature feature vector from the temperature time series data, the multidimensional temperature feature vector includes: the instantaneous temperature difference between the upper end cover and the lower end cover measuring point, the temperature change rate of the upper end cover and the lower end cover measuring point respectively, and the time series lag feature of the temperature response of the upper end cover and the lower end cover measuring point;Based on the multidimensional temperature feature vector, the current process phase is identified, and the process phase includes at least one of the pressurization stage, the pressure maintaining stage, the pressure relief stage and the cycle stage.

Description

Technical Field

[0001] This invention relates to the field of ultra-high pressure vessel monitoring technology, specifically to a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring. Background Technology

[0002] Extremely high-pressure process equipment such as isostatic presses, hot isostatic presses, and ultra-high-pressure reactors typically operate at pressures in the hundreds of megapascals or even gigapascals. Under such harsh conditions, directly installing temperature sensors inside the vessel shell presents several technical challenges: First, the sensor packaging structure must possess pressure resistance comparable to the vessel itself, resulting in extremely high manufacturing costs; second, the dynamic / static sealing structure at the lead wire perforation cannot guarantee long-term reliability under ultra-high-pressure cyclic loading, significantly increasing the risk of leakage; more critically, conventional temperature-sensitive elements are prone to piezoresistive shift, packaging structure collapse, and even functional failure under ultra-high static pressure, making long-term stable in-situ measurement difficult.

[0003] Currently, the industry commonly employs indirect monitoring strategies, inferring internal conditions by deploying offline built-in recorders or utilizing physical fields such as optics and acoustics. However, these methods all have significant limitations: offline built-in recorders can only acquire historical data after the fact, completely lacking the ability to obtain real-time data; while physical field measurement methods such as optics and acoustics are easily affected by ultra-high pressure media, making it difficult to guarantee measurement accuracy and stability, and the equipment is complex and costly. In summary, existing indirect monitoring strategies struggle to balance real-time performance, reliability, and cost-effectiveness. There is an urgent need to develop a non-invasive, highly reliable indirect sensing method for the internal thermal state of ultra-high pressure vessels to fill the technological gap in existing monitoring methods.

[0004] This section is intended to provide background or context for the embodiments of the invention set forth in the claims. The description herein is not an admission that it is prior art simply because it is included in this section. Summary of the Invention

[0005] To address at least one of the aforementioned problems in the prior art, embodiments of the present invention provide a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring.

[0006] This invention provides a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring. At least two temperature sensors are installed in the non-pressure area of ​​the upper end cover of the working cylinder of the ultra-high pressure vessel, and at least two temperature sensors are also installed in the non-pressure area of ​​the lower end cover. The method includes the following steps: S1. Synchronously collect the temperature time-series data of each of the temperature sensors; S2. Extract a multidimensional temperature feature vector from the temperature time series data. The multidimensional temperature feature vector includes: the instantaneous temperature difference between the upper and lower end cover measuring points, the temperature change rate of the upper and lower end cover measuring points respectively, and the time lag characteristics of the temperature response of the upper and lower end cover measuring points. S3. Based on the multidimensional temperature feature vector, identify the current process stage, which includes at least one of the pressurization stage, pressure holding stage, pressure release stage, and circulation stage.

[0007] In some embodiments, the instantaneous temperature difference ΔT(t) is calculated as follows: ΔT(t) = Tu(t) - Td(t), where Tu(t) is the average temperature of each measuring point on the upper end cover at time t, and Td(t) is the average temperature of each measuring point on the lower end cover at time t. The temperature change rate is calculated as follows: the average temperature sequence of each measuring point on the upper end cover is numerically differentiated to obtain the temperature change rate dTu / dt of the upper end cover measuring point; the average temperature sequence of each measuring point on the lower end cover is numerically differentiated to obtain the temperature change rate dTd / dt of the lower end cover measuring point. The time lag feature τ is calculated as follows: calculate the cross-correlation function Rud(τ) between the average temperature sequence of the upper end cover measuring point and the average temperature sequence of the lower end cover measuring point, and take the time offset that makes Rud(τ) reach its maximum value as the time lag feature τ.

[0008] In some embodiments, the multidimensional temperature feature vector F(t) is expressed as: F(t) = [Tu(t), Td(t), ΔT(t), dTu / dt, dTd / dt, τ], where Tu(t) and Td(t) are the average temperatures of the measuring points on the upper and lower end caps at time t, respectively.

[0009] In some embodiments, the method further includes: A heat conduction simulation model of the working cylinder, upper end cover, lower end cover and internal medium was established using multiphysics simulation software. The heat conduction simulation model is used to simulate the temperature response characteristics of the upper and lower end caps under different process stages, generating a simulation feature library for different process stages.

[0010] In some embodiments, identifying the current process stage based on the multidimensional temperature feature vector includes: Calculate the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library; The current process stage is identified based on the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library.

[0011] In some embodiments, the method further includes: Collect temperature data from the historical operation of the ultra-high pressure vessel, extract multi-dimensional temperature feature vectors for each time period, and generate process stage labels corresponding to each multi-dimensional temperature feature vector. A training sample set is established based on the multidimensional temperature feature vector and its process stage labels; A machine learning algorithm is used to train a classifier based on the training sample set to obtain a process stage discrimination model.

[0012] In some embodiments, identifying the current process stage based on the multidimensional temperature feature vector includes: The multidimensional temperature feature vector is input into the process stage discrimination model to obtain the current process stage output by the process stage discrimination model.

[0013] In some embodiments, the machine learning algorithm includes a support vector machine, a convolutional neural network, or a long short-term memory network.

[0014] In some embodiments, the method further includes: establishing a heat conduction simulation model of the working cylinder, upper end cover, lower end cover and internal medium using multiphysics simulation software; using the heat conduction simulation model to simulate the temperature response characteristics of the upper end cover and lower end cover under different process stages, and generating a simulation feature library for different process stages; The process of identifying the current process stage based on the multidimensional temperature feature vector includes: Calculate the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library, select the top K simulation feature vectors with the smallest Euclidean distance to the multidimensional temperature feature vector, count the frequency of occurrence of various process stages corresponding to the top K simulation feature vectors, and obtain the matching voting results, where K is a positive integer; The multidimensional temperature feature vector is input into the process stage discrimination model to obtain the posterior probability of each process stage. The current process stage is determined based on the matching voting results and the posterior probabilities of each process stage.

[0015] In some embodiments, determining the current process stage based on the matching voting results and the posterior probabilities of the various process stages includes: Calculate the fusion score for each process stage: S(c) = α V(c)+(1-α) PML(c|F), where α is the weighting coefficient, S(c) represents the fusion score of the process stage of category c, V(c) represents the frequency of the process stage of category c among the various process stages corresponding to the first K simulation feature vectors, and PML(c|F) represents the posterior probability of the process stage of category c. The process stage with the highest fusion score is selected as the current process stage.

[0016] This invention also provides a non-invasive temperature monitoring-based system for identifying the process stage of an ultra-high pressure vessel, comprising: The ultra-high pressure vessel body includes a working cylinder, an upper end cover, and a lower end cover; The temperature measurement module includes at least four temperature sensors, of which at least two are installed in the non-pressure area of ​​the upper end cover and at least two are installed in the non-pressure area of ​​the lower end cover; The data acquisition and processing unit, electrically connected to each of the temperature sensors, is used to: synchronously acquire the temperature time-series data of each of the temperature sensors; extract a multi-dimensional temperature feature vector from the temperature time-series data, the multi-dimensional temperature feature vector including: the instantaneous temperature difference between the measuring points of the upper and lower end caps, the temperature change rate of each measuring point of the upper and lower end caps, and the time lag characteristics of the temperature response of the measuring points of the upper and lower end caps; and identify the current process stage based on the multi-dimensional temperature feature vector, the process stage including at least one of the pressurization stage, pressure holding stage, pressure release stage, and circulation stage.

[0017] In some embodiments, the temperature sensor is a contact temperature sensor, including a PT100 platinum resistance thermometer or a K-type thermocouple.

[0018] In some embodiments, the temperature sensor is embedded inside the upper and lower end covers near the center of the end covers, and 5 to 20 mm from the inner surface of the end covers.

[0019] In some embodiments, the ultra-high pressure vessel includes a thermostatic press, a hot isostatic press, or an ultra-high pressure reactor.

[0020] This invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the program to implement the method described in any of the above embodiments.

[0021] This invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the above embodiments.

[0022] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the methods described in any of the above embodiments.

[0023] The proposed method for identifying process stages of ultra-high pressure vessels based on non-invasive temperature monitoring integrates temperature sensors into the non-pressurized areas of the upper and lower end covers. All measuring points are located outside the pressure-bearing cavity of the working cylinder, fundamentally avoiding the structural discontinuities and stress concentration problems caused by openings in the vessel shell in traditional invasive measurement methods. This eliminates potential failure risks such as high-pressure medium leakage and end cover ejection, achieving safe non-invasive monitoring while ensuring the integrity of the vessel structure. Addressing the technical bottleneck of the thick walls and significant thermal inertia of ultra-high pressure vessels leading to sluggish response in traditional external wall temperature measurement, this invention constructs a multi-dimensional temperature feature tensor encompassing spatial differences (instantaneous temperature difference ΔT), temporal gradients (rate of change dT / dt), and response order (time lag τ) by deploying multiple measuring points (at least two at each end) on the upper and lower end covers. This effectively captures the subtle thermal response characteristics corresponding to different process stages (e.g., pressurization, pressure holding, pressure release, and circulation), improving the real-time performance and accuracy of process stage identification. Because the temperature measurement points are located in normal or low-pressure environments, the sensor does not need to withstand ultra-high pressure loads and can directly use conventional industrial-grade temperature sensors, completely eliminating the need for expensive high-pressure resistant packaging and special sealed lead structures, thus significantly reducing hardware costs. At the same time, the sensing element avoids the interference of high-pressure media on measurement accuracy (such as the piezoresistive effect) and the risk of physical damage, ensuring measurement stability and reliability under long-term service conditions. Furthermore, it is not dependent on specific container sizes or media types, possessing good geometric and physical universality, and has broad engineering application prospects and promotional value. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart illustrating a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring, as provided in an embodiment of the present invention.

[0025] Figure 2 This is a partial flowchart illustrating a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring, as provided in an embodiment of the present invention.

[0026] Figure 3 This is a partial flowchart illustrating a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring, as provided in an embodiment of the present invention.

[0027] Figure 4This is a partial flowchart illustrating a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring, as provided in an embodiment of the present invention.

[0028] Figure 5 This is a partial flowchart illustrating a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring, as provided in an embodiment of the present invention.

[0029] Figure 6 This is a partial flowchart illustrating a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring, as provided in an embodiment of the present invention.

[0030] Figure 7 This is a schematic diagram of a process stage identification system for ultra-high pressure vessels based on non-invasive temperature monitoring, provided in an embodiment of the present invention.

[0031] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0032] Detailed reference is made to exemplary embodiments, examples of which are illustrated in the accompanying drawings. The following description refers to the accompanying drawings, wherein the same reference numerals in different drawings denote the same or similar elements unless otherwise stated. In the following description of exemplary embodiments, the described embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with several aspects of the invention as described in the appended claims.

[0033] The terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. In this invention and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein is intended to represent and include any or all possible combinations of one or more of the associated listed items.

[0034] It should be understood that although the terms “first,” “second,” “third,” etc., may be used herein to describe various types of information, such information should not be limited by these terms. These terms are only used to distinguish the types of information. For example, without departing from the scope of the invention, first information may be referred to as second information, and similarly, second information may be referred to as first information. The term “if” as used herein may be understood, depending on the context, to mean “at the time,” “at this moment,” or “in response to a judgment.”

[0035] To address the problems existing in the prior art, this invention provides a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring.

[0036] Figure 1 This is a flowchart illustrating a method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring, as provided in an embodiment of the present invention. At least two temperature sensors are installed in the non-pressure area of ​​the upper end cover of the working cylinder of the ultra-high pressure vessel, and at least two temperature sensors are also installed in the non-pressure area of ​​the lower end cover; as shown... Figure 1 As shown, the method includes the following steps: S1. Synchronously collect the temperature time-series data of each of the temperature sensors; In step S1, during the operation of the ultra-high pressure vessel, temperature sensor data installed in the non-pressurized areas of the upper and lower end covers can be synchronously acquired using a data acquisition device. The acquisition frequency can be set to ≥10Hz to form a continuous temperature time-series data stream. During the temperature data acquisition process, the acquisition timestamp is automatically recorded to ensure the timeliness and integrity of the data.

[0037] S2. Extract a multidimensional temperature feature vector from the temperature time series data. The multidimensional temperature feature vector includes: the instantaneous temperature difference between the upper and lower end cover measuring points, the temperature change rate of the upper and lower end cover measuring points respectively, and the time lag characteristics of the temperature response of the upper and lower end cover measuring points. In step S2, the instantaneous temperature difference ΔT(t) is calculated using the formula: ΔT(t) = Tu(t) - Td(t), where Tu(t) is the average temperature of each measuring point on the upper end cover at time t, and Td(t) is the average temperature of each measuring point on the lower end cover at time t. First, the average temperature value Tu(t) of all temperature sensors installed on the upper end cover at time t is calculated, then the average temperature value Td(t) of all temperature sensors installed on the lower end cover at time t is calculated, and finally the difference between the two is calculated to obtain the instantaneous temperature difference ΔT(t).

[0038] The temperature change rate is calculated as follows: The average temperature sequence of each measuring point on the upper end cover is numerically differentiated to obtain the temperature change rate dTu / dt of the upper end cover measuring points; the average temperature sequence of each measuring point on the lower end cover is numerically differentiated to obtain the temperature change rate dTd / dt of the lower end cover measuring points; the numerical differentiation can be performed using the central difference method, and the calculation formula is dTu / dt = (Tu(t+Δt) - Tu(t-Δt)) / (2 Δt), where Δt is the sampling time interval; similarly, dTd / dt = (Td(t+Δt) - Td(t-Δt)) / (2 Δt).

[0039] The time lag feature τ is calculated as follows: The cross-correlation function Rud(τ) between the average temperature sequences of the upper and lower end cap measuring points is calculated, and the time offset that makes Rud(τ) reach its maximum value is taken as the time lag feature τ. The formula for calculating the cross-correlation function Rud(τ) is: Rud(τ) = Σ[Tu(t)Td(t+τ)], where τ is the time offset. By iterating through different τ values, the τ that makes Rud(τ) reach its maximum value is found, which is the desired time lag feature.

[0040] In addition to the four feature data mentioned above, the multidimensional temperature feature vector F(t) can also include the average temperature data of the measuring points on the upper and lower end caps at time t. Specifically, the multidimensional temperature feature vector F(t) can be expressed as: F(t) = [Tu(t), Td(t), ΔT(t), dTu / dt, dTd / dt, τ], where Tu(t) and Td(t) are the average temperatures of the measuring points on the upper and lower end caps at time t, respectively. Thus, a six-dimensional feature vector F(t) can be obtained at each time t for subsequent process stage identification.

[0041] S3. Based on the multidimensional temperature feature vector, identify the current process stage, which includes at least one of the pressurization stage, pressure holding stage, pressure release stage, and circulation stage.

[0042] In step S3, during the identification process, at least one of the following methods can be used to determine the process stage: physical model-based methods, machine learning-based methods, etc.

[0043] The method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring provided in this invention integrates temperature sensors into the non-pressurized areas of the upper and lower end caps. All measuring points are located outside the pressure-bearing cavity of the working cylinder, fundamentally avoiding the structural discontinuities and stress concentration problems caused by openings in the vessel cylinder in traditional invasive measurement methods. It also eliminates potential failure risks such as high-pressure medium leakage and end cap ejection, achieving safe non-invasive monitoring while ensuring the structural integrity of the vessel.

[0044] To address the technical bottleneck of traditional external wall temperature measurement lag caused by the thick walls and significant thermal inertia of ultra-high pressure vessels, a multidimensional temperature feature tensor was constructed by deploying multiple measurement points (at least two at each end) on the upper and lower end covers. This tensor encompasses spatial differences (instantaneous temperature difference ΔT), temporal gradients (rate of change dT / dt), and response order (time lag τ). It can effectively capture the subtle thermal response characteristics corresponding to different process stages (e.g., pressurization, pressure holding, pressure release, and circulation), improving the real-time performance and accuracy of process stage identification.

[0045] Because the temperature measurement points are located in normal or low-pressure environments, the sensor does not need to withstand ultra-high pressure loads and can directly use conventional industrial-grade temperature sensors, completely eliminating the need for expensive high-pressure resistant packaging and special sealed lead structures, thus significantly reducing hardware costs. At the same time, the sensing element avoids the interference of high-pressure media on measurement accuracy (such as the piezoresistive effect) and the risk of physical damage, ensuring measurement stability and reliability under long-term service conditions. Furthermore, it is not dependent on specific container sizes or media types, possessing good geometric and physical universality, and has broad engineering application prospects and promotional value.

[0046] like Figure 2 As shown, in some embodiments, the method further includes: S01. Establish a heat conduction simulation model of the working cylinder, upper end cover, lower end cover and internal medium using multiphysics simulation software; S02. Use the heat conduction simulation model to simulate the temperature response characteristics of the upper and lower end caps under different process stages, and generate a simulation feature library for different process stages.

[0047] Specifically, a heat conduction simulation model of the working cylinder, upper end cover, lower end cover, and internal medium is established using multiphysics simulation software. This model considers factors such as the thermophysical properties and geometry of the container material, and the flow and heat transfer characteristics of the internal medium. This heat conduction simulation model is used to simulate the temperature response characteristics of the upper and lower end covers under different process stages, generating simulation feature libraries for different process stages. During actual operation, the real-time extracted multidimensional temperature feature vectors are compared with the simulation feature vectors in the aforementioned simulation feature library to identify the current process stage. The identification results can also be transmitted to the industrial control system to achieve real-time monitoring and closed-loop control of the process.

[0048] like Figure 3 As shown, in some embodiments, when having simulation feature libraries for different process stages, identifying the current process stage based on the multidimensional temperature feature vector includes: S31. Calculate the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library; S32. Based on the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library, identify the current process stage.

[0049] Specifically, the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library is calculated. The process stage corresponding to the simulation feature vector with the smallest distance to the multidimensional temperature feature vector can be selected as the current process stage. Alternatively, the top K simulation feature vectors with the smallest distance (e.g., K=5) can be selected, and the frequency of each type of process stage in these simulation feature vectors can be counted. The process stage with the highest frequency can be selected as the current process stage.

[0050] like Figure 4 As shown, in some embodiments, the method further includes: S03. Collect temperature data from the historical operation of the ultra-high pressure vessel, extract multi-dimensional temperature feature vectors for each time period, and generate process stage labels corresponding to each multi-dimensional temperature feature vector. S04. Establish a training sample set based on the multidimensional temperature feature vector and its process stage labels; S05. A machine learning algorithm is used to train a classifier based on the training sample set to obtain a process stage discrimination model.

[0051] Specifically, multiple sets of temperature data from the actual equipment of the ultra-high pressure vessel during its historical operation are collected. Process experts label each time period with process stage tags (pressurization, pressure holding, pressure release, and circulation). Feature extraction is performed on the labeled data to obtain a training sample set. A machine learning algorithm is then used to train a classifier based on this training sample set to obtain a process stage discrimination model. The machine learning algorithm includes support vector machines, convolutional neural networks, or long short-term memory networks.

[0052] In some embodiments, when the process stage discrimination model is present, identifying the current process stage based on the multidimensional temperature feature vector may include: inputting the multidimensional temperature feature vector into the process stage discrimination model to obtain the current process stage output by the process stage discrimination model.

[0053] Specifically, during actual operation, the real-time extracted temperature feature vector is input into the process stage discrimination model, and the model outputs the current process stage. Furthermore, the model output can be transmitted to the industrial control system to achieve real-time monitoring and closed-loop control of the process.

[0054] like Figure 5 As shown, in some embodiments, when both simulation feature libraries for different process stages and the process stage discrimination model are available, a discrimination mechanism that integrates a prior feature library (simulation feature library) constructed using multiphysics simulation with a machine learning classification model (process stage discrimination model) is employed. Specifically, identifying the current process stage based on the multidimensional temperature feature vector may include: S33. Calculate the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library, select the top K simulation feature vectors with the smallest Euclidean distance to the multidimensional temperature feature vector, count the frequency of occurrence of various process stages corresponding to the top K simulation feature vectors, and obtain the matching voting results, where K is a positive integer; S34. Input the multidimensional temperature feature vector into the process stage discrimination model to obtain the posterior probability of each type of process stage; S35. Based on the matching voting results and the posterior probabilities of the various process stages, determine the current process stage.

[0055] Specifically, during real-time operation, the multidimensional temperature feature vector F(t) at the current moment is processed in the following two steps: (1) Feature matching: Calculate the Euclidean distance between F(t) and each simulation feature vector in the simulation feature library L, select the K simulation feature vectors with the smallest distance (e.g., K=5 or 6), count the frequency of each process stage in these simulation feature vectors, and obtain the matching voting result V.

[0056] (2) Machine learning classification: Input F(t) into the trained process stage discrimination model (SVM classifier) ​​to obtain the posterior probability PML(c|F) of each category.

[0057] Finally, the current process stage is determined by fusion of the matching voting result V and the posterior probability PML(c|F) of each category.

[0058] like Figure 6 As shown, in some embodiments, determining the current process stage based on the matching voting results and the posterior probabilities of various process stages includes: S351. Calculate the fusion score for each type of process stage: S(c) = α V(c)+(1-α) PML(c|F), where α is the weighting coefficient, for example, α=0.4, S(c) represents the fusion score of the process stage of category c, V(c) represents the frequency of the process stage of category c among the various process stages corresponding to the first K simulation feature vectors, and PML(c|F) represents the posterior probability of the process stage of category c. S352. Select the process stage with the highest fusion score as the current process stage.

[0059] Furthermore, by adjusting the parameters of the heat conduction simulation model or by performing transfer learning on the process stage discrimination model based on new operating data, it is possible to quickly adapt to ultra-high pressure process equipment of different specifications (such as thermal isostatic press, hot isostatic press, and ultra-high pressure reactor).

[0060] Based on the same inventive concept, embodiments of the present invention also provide a process stage identification system for ultra-high pressure vessels based on non-invasive temperature monitoring.

[0061] Figure 7 This is a schematic diagram of a non-invasive temperature monitoring-based process stage identification system for ultra-high pressure vessels provided in an embodiment of the present invention. Figure 7 As shown in the figure, an embodiment of the present invention provides a process stage identification system for ultra-high pressure vessels based on non-invasive temperature monitoring, comprising: The ultra-high pressure vessel body 1 includes a working cylinder, an upper end cover, and a lower end cover; Temperature measurement module 2 includes at least four temperature sensors, of which at least two are installed in the non-pressure area of ​​the upper end cover and at least two are installed in the non-pressure area of ​​the lower end cover; The data acquisition and processing unit 3 is electrically connected to each of the temperature sensors and is used to: synchronously acquire the temperature time-series data of each of the temperature sensors; extract a multi-dimensional temperature feature vector from the temperature time-series data, the multi-dimensional temperature feature vector including: the instantaneous temperature difference between the upper and lower end cap measuring points, the temperature change rate of each measuring point of the upper and lower end caps, and the time lag characteristics of the temperature response of the upper and lower end cap measuring points; and identify the current process stage based on the multi-dimensional temperature feature vector, the process stage including at least one of the pressurization stage, pressure holding stage, pressure release stage, and circulation stage.

[0062] Specifically, the ultra-high pressure vessel includes a working cylinder, an upper end cover module, and a lower end cover module that cooperate with it. The upper end cover module may include an upper end cover, a circular metal connector, and a hydraulic jacking cylinder. The upper end cover is sealed and installed at the upper port of the working cylinder. The hydraulic jacking cylinder is connected to the upper end cover via the circular metal connector and is used to drive the upper end cover to open or close. The lower end cover module includes a lower end cover and a base. The lower end cover is sealed and installed at the lower port of the working cylinder and fixed to the base. The temperature measurement module includes at least four temperature sensors, of which at least two temperature sensors are installed on the outer end face of the upper end cover or embedded inside the upper end cover, and at least two temperature sensors are installed on the outer end face of the lower end cover or embedded inside the lower end cover; all temperature sensors are located outside the pressure chamber of the working cylinder and in contact with the non-pressure area, and are used to collect temperature response data of the upper end cover and the lower end cover in real time. The data acquisition and processing unit is electrically connected to each temperature sensor to synchronously acquire multi-channel temperature time-series data and perform preprocessing such as filtering and normalization on the data. Based on the preprocessed data, it calculates the instantaneous temperature difference ΔT between the measuring points of the upper and lower end caps, the rate of change of temperature at each measuring point over time dT / dt, and the time lag characteristic τ of the temperature response of the upper and lower end caps, and constructs a multi-dimensional temperature feature vector. Based on the multi-dimensional temperature feature vector, it identifies the current process stage, including but not limited to the pressurization stage, pressure holding stage, pressure release stage, and cyclic pressing stage.

[0063] The ultra-high pressure vessel process stage identification system based on non-invasive temperature monitoring provided in this invention integrates temperature sensors into the non-pressurized areas of the upper and lower end caps. All measuring points are located outside the pressure-bearing cavity of the working cylinder, fundamentally avoiding the structural discontinuity and stress concentration problems caused by the opening of the vessel cylinder in traditional invasive measurement methods. It also eliminates potential failure risks such as high-pressure medium leakage and end cap ejection, and realizes safe non-invasive monitoring while ensuring the integrity of the vessel structure.

[0064] To address the technical bottleneck of traditional external wall temperature measurement lag caused by the thick walls and significant thermal inertia of ultra-high pressure vessels, a multidimensional temperature feature tensor was constructed by deploying multiple measurement points (at least two at each end) on the upper and lower end covers. This tensor encompasses spatial differences (instantaneous temperature difference ΔT), temporal gradients (rate of change dT / dt), and response order (time lag τ). It can effectively capture the subtle thermal response characteristics corresponding to different process stages (e.g., pressurization, pressure holding, pressure release, and circulation), improving the real-time performance and accuracy of process stage identification.

[0065] Because the temperature measurement points are located in normal or low-pressure environments, the sensor does not need to withstand ultra-high pressure loads and can directly use conventional industrial-grade temperature sensors, completely eliminating the need for expensive high-pressure resistant packaging and special sealed lead structures, thus significantly reducing hardware costs. At the same time, the sensing element avoids the interference of high-pressure media on measurement accuracy (such as the piezoresistive effect) and the risk of physical damage, ensuring measurement stability and reliability under long-term service conditions. Furthermore, it is not dependent on specific container sizes or media types, possessing good geometric and physical universality, and has broad engineering application prospects and promotional value.

[0066] In some embodiments, the temperature sensor is a contact temperature sensor, including a PT100 platinum resistance thermometer or a K-type thermocouple.

[0067] In some embodiments, the temperature sensor is embedded inside the upper and lower end covers near the center of the end covers, and 5 to 20 mm from the inner surface of the end covers.

[0068] In some embodiments, the ultra-high pressure vessel includes a thermostatic press, a hot isostatic press, or an ultra-high pressure reactor.

[0069] To better understand the present invention, the following detailed description of the method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring provided by the present invention is provided through a specific embodiment.

[0070] This embodiment proposes a non-invasive, low-cost, and easy-to-implement method and system for determining the internal process state of an ultra-high pressure vessel. By simply placing a small number of temperature measuring points outside the vessel or in non-pressurized areas, the method utilizes the temperature difference characteristics between multiple points to inversely deduce the process stage occurring inside the cavity.

[0071] This embodiment proposes a non-invasive temperature monitoring-based process stage identification system for ultra-high pressure vessels, comprising: The ultra-high pressure vessel body includes a working cylinder (shell), an upper end cover module, and a lower end cover module. The upper end cover module includes an upper end cover, a circular metal connector, and a hydraulic jacking cylinder. The upper end cover is opened and closed by the hydraulic jacking cylinder and employs a self-tightening sealing structure to seal against the upper port of the shell. The lower end cover module includes a lower end cover and a base. The lower end cover is fixedly installed at the lower port of the shell and rigidly connected to the base.

[0072] Temperature measurement module: Includes four high-precision temperature sensors (e.g., PT100 platinum resistance thermometers or K-type thermocouples). Two sensors are embedded inside the upper end cover near the center of the end cover, at a distance of 5-20 mm from the inner surface of the upper end cover (the surface in contact with the medium). The other two sensors are installed inside the lower end cover in the same manner. All sensors are located outside the pressure chamber of the cylinder and do not bear ultra-high pressure loads. The sensor leads are led out through the normal pressure lead holes on the end cover, eliminating the need for a high-pressure sealing structure.

[0073] Data acquisition and processing unit: Electrically connected to four temperature sensors, using a synchronous acquisition module (sampling frequency ≥10 Hz, resolution ≥16 bits), with built-in signal conditioning circuit (filtering, amplification) and microprocessor, used for real-time acquisition and processing of temperature data and execution of discrimination algorithms.

[0074] Industrial control interface: The data acquisition and processing unit connects to the PLC / DCS system via industrial Ethernet or fieldbus, and outputs the process stage judgment results in real time for use in closed-loop control or safety interlocking.

[0075] In actual operation, the data acquisition and processing unit synchronously acquires the temperature values ​​of four measuring points at a fixed sampling frequency (e.g., 20 Hz), denoted as: Temperature measurements at the top cover: Tu1, Tu2; Temperatures at measuring points on the lower end cover: Td1, Td2; The original temperature sequence is filtered using a moving average (window length of 0.5 seconds) to eliminate random noise and outlier removal (e.g., using the 3σ criterion). The preprocessed temperature data is then used for subsequent feature extraction.

[0076] In this embodiment, three types of features are extracted from the preprocessed temperature sequence to construct a multi-dimensional feature vector for process stage discrimination.

[0077] 1. Spatial characteristics: Instantaneous temperature difference ΔT(t) between the upper and lower end caps

[0078] Average temperature of the upper cover: Tu(t) = [Tu1(t) + Tu2t)] / 2; Average temperature of the lower end cap: Td(t) = [Td1(t) + Td2(t)] / 2; The instantaneous temperature difference is: ΔT(t) = Tu(t) - Td(t); This feature reflects the spatial temperature difference between the upper and lower end caps, which is closely related to the heat distribution of the internal medium.

[0079] 2. Time characteristics: Rate of temperature change dT / dt

[0080] Numerical differentiation of the average temperature sequence of the upper end cap (using central difference or Savitzky-Golay filtering for smoothing and differentiation) yields the rate of change of the upper end cap temperature: dTu / dt≈[Tu(t+Δt)-Tu(t-Δt)] / (2Δt).

[0081] Similarly, the temperature change rate dTd / dt of the lower end cap can be obtained. The temperature change rate characterizes the transient response rate of the end cap temperature and can reflect the drastic degree of change in the internal medium state.

[0082] 3. Phase characteristics: Response time lag τ

[0083] Because the upper and lower end caps are located at different positions relative to the heat source of the internal medium, their temperature responses exhibit phase differences. The cross-correlation function of the average temperature sequences of the upper and lower end caps is calculated as follows: Rud(τ) = Tu(t)Td(t+τ)dt.

[0084] The time offset that makes Rud(τ) reach its maximum value is taken as the response timing lag τ, that is: τ=argmax Rud(τ).

[0085] A positive τ indicates that the upper cap's response leads the lower cap's response, while a negative τ indicates the opposite. This characteristic reflects the direction and speed of heat flow within the container.

[0086] 4. Feature Vector Construction

[0087] Combining the above features with the original temperature value forms a multidimensional feature vector for the current time t: F(t) = [Tu(t), Td(t), ΔT(t), dTu / dt, dTd / dt,τ]; To enhance the robustness of the discrimination, feature vectors from multiple consecutive time points can be stacked to form a feature tensor (e.g., taking a data window from the past 5 seconds) and input into the subsequent discrimination model.

[0088] A thermal conductivity simulation model of an ultra-high pressure vessel with identical geometric dimensions and material properties to the actual vessel was established using finite element method (FEM) software. In the model, actual thermal contact resistance was set between the upper and lower end caps and the cylinder. The thermal properties (thermal conductivity, specific heat capacity) of the internal medium (such as water, oil, or gas) as a function of pressure were input from literature or experimental data.

[0089] Simulation of four typical process stages: Pressurization stage: The internal pressure rises linearly from atmospheric pressure to working pressure, taking into account the heat effect of compression (the temperature of the medium increases).

[0090] Pressure holding stage: The pressure remains constant for several tens of minutes, while the temperature of the medium changes due to natural cooling or the action of the heating system.

[0091] Depressurization phase: The pressure drops rapidly to atmospheric pressure, taking into account the heat absorption effect of medium expansion.

[0092] Circulation stage: The process involves preheating the inside of the cylinder.

[0093] For each process stage, the simulation outputs the temperature response time series data at specified measuring points on the upper and lower end caps, and generates a simulation feature vector F and its corresponding process stage label using the same feature extraction method as the actual measurement, thus forming a simulation feature library L.

[0094] Multiple sets of temperature data from the actual equipment during historical operation were collected, and process experts labeled each time period with process stage tags (pressurization, pressure holding, pressure release, circulation). Feature extraction was performed on the labeled data to obtain a training sample set.

[0095] During real-time execution, the feature vector F(t) at the current time step is processed in the following two steps: (1) Feature matching: Calculate the Euclidean distance between F(t) and each feature vector in the simulation feature library L, select the K samples with the smallest distance (K=5), count the frequency of each process stage in these samples, and obtain the matching voting result V.

[0096] (2) Machine learning classification: Input F(t) into the trained SVM classifier to obtain the posterior probability PML(c|F) of each category.

[0097] Finally, a weighted voting mechanism is used for decision fusion. The fusion score S(c) for the c-th process stage is calculated as follows: S(c) = α V+(1-α) PML(c|F); Where α is the weighting coefficient (in this embodiment, α = 0.4, adjusted according to the performance of the validation set). The category with the highest score is selected as the discrimination result for the current process stage.

[0098] The embodiments of the present invention are not limited to the above examples, and can be adjusted as follows according to actual needs: The number of temperature sensors can be increased to three or more per end to improve measurement redundancy and feature robustness.

[0099] Feature extraction can utilize more derived features, such as the second derivative of temperature and wavelet packet energy.

[0100] The discrimination model can adopt an end-to-end deep learning approach, directly inputting multi-channel temperature time series data into a convolutional neural network (CNN) or LSTM to achieve automatic feature extraction and classification.

[0101] For certain types of equipment, if there is sufficient historical data, the simulation feature library can be omitted, and only the data-driven model can be used; conversely, if there is insufficient historical data, the simulation feature library can be relied upon for matching.

[0102] The weighting coefficient α can be adaptively adjusted according to the actual data, such as by using Bayesian model averaging.

[0103] Figure 8 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, as shown below. Figure 8 As shown, the electronic device may include a processor 301, a communications interface 302, a memory 303, and a communication bus 304, wherein the processor 301, the communications interface 302, and the memory 303 communicate with each other via the communication bus 304. The processor 301 may call logical instructions in the memory 303 to execute the methods described in any of the above embodiments.

[0104] Furthermore, the logical instructions in the aforementioned memory 303 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0105] This embodiment of the invention provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer can perform the methods provided in the above-described method embodiments.

[0106] This embodiment provides a computer-readable storage medium storing a computer program that causes the computer to perform the methods provided in the above-described method embodiments.

[0107] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0108] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0109] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0110] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0111] In the description of this specification, the references to terms such as "an embodiment," "a specific embodiment," "some embodiments," "for example," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0112] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring, characterized in that, At least two temperature sensors are installed in the non-pressure area of ​​the upper end cover of the working cylinder of the ultra-high pressure vessel, and at least two temperature sensors are also installed in the non-pressure area of ​​the lower end cover; the method includes the following steps: S1. Synchronously collect the temperature time-series data of each of the temperature sensors; S2. Extract a multidimensional temperature feature vector from the temperature time series data. The multidimensional temperature feature vector includes: the instantaneous temperature difference between the upper and lower end cap measuring points, the temperature change rate of each measuring point on the upper and lower end caps, and the time lag feature of the temperature response of the upper and lower end cap measuring points. The time lag feature τ is calculated as follows: calculate the cross-correlation function Rud(τ) between the average temperature sequence of the upper end cap measuring point and the average temperature sequence of the lower end cap measuring point, and take the time offset that makes Rud(τ) reach its maximum value as the time lag feature τ. S3. Based on the multidimensional temperature feature vector, identify the current process stage, wherein the process stage includes at least one of the pressurization stage, pressure holding stage, pressure release stage and circulation stage; The method further includes: establishing a heat conduction simulation model of the working cylinder, upper end cover, lower end cover and internal medium using multiphysics simulation software; using the heat conduction simulation model to simulate the temperature response characteristics of the upper end cover and lower end cover under different process stages, and generating a simulation feature library for different process stages; The step of identifying the current process stage based on the multidimensional temperature feature vector includes: calculating the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library; and identifying the current process stage based on the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library.

2. The method according to claim 1, characterized in that, The instantaneous temperature difference ΔT(t) is calculated as follows: ΔT(t) = Tu(t) - Td(t), where Tu(t) is the average temperature of each measuring point on the upper end cover at time t, and Td(t) is the average temperature of each measuring point on the lower end cover at time t. The temperature change rate is calculated as follows: the average temperature sequence of each measuring point on the upper end cover is numerically differentiated to obtain the temperature change rate dTu / dt of the upper end cover measuring points; the average temperature sequence of each measuring point on the lower end cover is numerically differentiated to obtain the temperature change rate dTd / dt of the lower end cover measuring points.

3. The method according to claim 2, characterized in that, The multidimensional temperature feature vector F(t) is expressed as: F(t) = [Tu(t), Td(t), ΔT(t), dTu / dt, dTd / dt, τ], where Tu(t) and Td(t) are the average temperatures of the measuring points on the upper and lower end caps at time t, respectively.

4. A method for identifying the process stage of an ultra-high pressure vessel based on non-invasive temperature monitoring, characterized in that, At least two temperature sensors are installed in the non-pressure area of ​​the upper end cover of the working cylinder of the ultra-high pressure vessel, and at least two temperature sensors are also installed in the non-pressure area of ​​the lower end cover; the method includes the following steps: S1. Synchronously collect the temperature time-series data of each of the temperature sensors; S2. Extract a multidimensional temperature feature vector from the temperature time series data. The multidimensional temperature feature vector includes: the instantaneous temperature difference between the upper and lower end cap measuring points, the temperature change rate of each measuring point on the upper and lower end caps, and the time lag feature of the temperature response of the upper and lower end cap measuring points. The time lag feature τ is calculated as follows: calculate the cross-correlation function Rud(τ) between the average temperature sequence of the upper end cap measuring point and the average temperature sequence of the lower end cap measuring point, and take the time offset that makes Rud(τ) reach its maximum value as the time lag feature τ. S3. Based on the multidimensional temperature feature vector, identify the current process stage, wherein the process stage includes at least one of the pressurization stage, pressure holding stage, pressure release stage and circulation stage; The method further includes: collecting temperature data from the historical operation of the ultra-high pressure vessel, extracting multi-dimensional temperature feature vectors for each time period, and generating process stage labels corresponding to each multi-dimensional temperature feature vector; establishing a training sample set based on the multi-dimensional temperature feature vectors and their process stage labels; and using a machine learning algorithm to train a classifier based on the training sample set to obtain a process stage discrimination model. The step of identifying the current process stage based on the multidimensional temperature feature vector includes: inputting the multidimensional temperature feature vector into the process stage discrimination model to obtain the current process stage output by the process stage discrimination model.

5. The method according to claim 4, characterized in that, The machine learning algorithms include support vector machines, convolutional neural networks, or long short-term memory networks.

6. The method according to claim 4, characterized in that, The method further includes: establishing a heat conduction simulation model of the working cylinder, upper end cover, lower end cover and internal medium using multiphysics simulation software; using the heat conduction simulation model to simulate the temperature response characteristics of the upper end cover and lower end cover under different process stages, and generating a simulation feature library for different process stages; The process of identifying the current process stage based on the multidimensional temperature feature vector includes: Calculate the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library, select the top K simulation feature vectors with the smallest Euclidean distance to the multidimensional temperature feature vector, count the frequency of occurrence of various process stages corresponding to the top K simulation feature vectors, and obtain the matching voting results, where K is a positive integer; The multidimensional temperature feature vector is input into the process stage discrimination model to obtain the posterior probability of each process stage. The current process stage is determined based on the matching voting results and the posterior probabilities of each process stage.

7. The method according to claim 6, characterized in that, The process of determining the current process stage based on the matching voting results and the posterior probabilities of various process stages includes: Calculate the fusion score for each process stage: S(c) = α V(c)+(1-α) PML(c|F), where α is the weighting coefficient, S(c) represents the fusion score of the process stage of category c, V(c) represents the frequency of the process stage of category c among the various process stages corresponding to the first K simulation feature vectors, and PML(c|F) represents the posterior probability of the process stage of category c. The process stage with the highest fusion score is selected as the current process stage.

8. A process stage identification system for ultra-high pressure vessels based on non-invasive temperature monitoring, characterized in that, include: The ultra-high pressure vessel body includes a working cylinder, an upper end cover, and a lower end cover; The temperature measurement module includes at least four temperature sensors, of which at least two are installed in the non-pressure area of ​​the upper end cover and at least two are installed in the non-pressure area of ​​the lower end cover; A data acquisition and processing unit, electrically connected to each of the temperature sensors, is used to: synchronously acquire temperature time-series data from each of the temperature sensors; extract a multi-dimensional temperature feature vector from the temperature time-series data, the multi-dimensional temperature feature vector including: the instantaneous temperature difference between the upper and lower end cap measuring points, the temperature change rate of each measuring point on the upper and lower end caps, and the time lag characteristics of the temperature response at the upper and lower end cap measuring points; and identify the current process stage based on the multi-dimensional temperature feature vector, the process stage including at least one of a pressurization stage, a pressure holding stage, a pressure release stage, and a circulation stage. The time lag feature τ is calculated as follows: the cross-correlation function Rud(τ) between the average temperature sequence of the upper end cap measuring point and the average temperature sequence of the lower end cap measuring point is calculated, and the time offset that makes Rud(τ) reach its maximum value is taken as the time lag feature τ. The step of identifying the current process stage based on the multidimensional temperature feature vector includes: calculating the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library; identifying the current process stage based on the Euclidean distance between the multidimensional temperature feature vector and each simulation feature vector in the simulation feature library; wherein, the process of establishing the simulation feature library is as follows: establishing a heat conduction simulation model of the working cylinder, upper end cover, lower end cover and internal medium using multiphysics simulation software; using the heat conduction simulation model to simulate the temperature response characteristics of the upper end cover and lower end cover under different process stages, generating simulation feature libraries for different process stages.

9. The system according to claim 8, characterized in that, The temperature sensor is a contact temperature sensor, including a PT100 platinum resistance thermometer or a K-type thermocouple.

10. The system according to claim 8, characterized in that, The temperature sensor is embedded inside the upper and lower end covers, near the center of the end cover, and 5-20 mm from the inner surface of the end cover.

11. The system according to claim 8, characterized in that, The ultra-high pressure vessel includes a thermostatic press, a hot isostatic press, or an ultra-high pressure reactor.

12. A process stage identification system for ultra-high pressure vessels based on non-invasive temperature monitoring, characterized in that, include: The ultra-high pressure vessel body includes a working cylinder, an upper end cover, and a lower end cover; The temperature measurement module includes at least four temperature sensors, of which at least two are installed in the non-pressure area of ​​the upper end cover and at least two are installed in the non-pressure area of ​​the lower end cover; A data acquisition and processing unit, electrically connected to each of the temperature sensors, is used to: synchronously acquire temperature time-series data from each of the temperature sensors; extract a multi-dimensional temperature feature vector from the temperature time-series data, the multi-dimensional temperature feature vector including: the instantaneous temperature difference between the upper and lower end cap measuring points, the temperature change rate of each measuring point on the upper and lower end caps, and the time lag characteristics of the temperature response at the upper and lower end cap measuring points; and identify the current process stage based on the multi-dimensional temperature feature vector, the process stage including at least one of a pressurization stage, a pressure holding stage, a pressure release stage, and a circulation stage. The time lag feature τ is calculated as follows: the cross-correlation function Rud(τ) between the average temperature sequence of the upper end cap measuring point and the average temperature sequence of the lower end cap measuring point is calculated, and the time offset that makes Rud(τ) reach its maximum value is taken as the time lag feature τ. The step of identifying the current process stage based on the multidimensional temperature feature vector includes: inputting the multidimensional temperature feature vector into a process stage discrimination model to obtain the current process stage output by the process stage discrimination model; wherein, the generation process of the process stage discrimination model is as follows: collecting temperature data from the historical operation of the ultra-high pressure vessel, extracting multidimensional temperature feature vectors for each time period and generating process stage labels corresponding to each multidimensional temperature feature vector; establishing a training sample set based on the multidimensional temperature feature vectors and their process stage labels; and training a classifier using a machine learning algorithm based on the training sample set to obtain the process stage discrimination model.

13. The system according to claim 12, characterized in that, The temperature sensor is a contact temperature sensor, including a PT100 platinum resistance thermometer or a K-type thermocouple.

14. The system according to claim 12, characterized in that, The temperature sensor is embedded inside the upper and lower end covers, near the center of the end cover, and 5-20 mm from the inner surface of the end cover.

15. The system according to claim 12, characterized in that, The ultra-high pressure vessel includes a thermostatic press, a hot isostatic press, or an ultra-high pressure reactor.

Images