Chromatograph cleaning method and system, intelligent terminal and storage medium

Abstract

The invention relates to the technical field of data cleaning, and discloses a chromatographic instrument cleaning method and system, an intelligent terminal and a storage medium, and the method comprises the following steps: carrying out structured analysis on multi-dimensional operation trace data and historical cleaning cases of a chromatographic instrument to obtain pollution characteristic data of the chromatographic instrument; performing topology construction on the cleaning state features in the historical cleaning cases to obtain a cleaning knowledge graph; performing optimal path deduction on the cleaning knowledge graph to obtain a personalized cleaning track; performing decision deduction on the personalized cleaning track to obtain an initial cleaning scheme, and performing prospective planning on the initial cleaning scheme to obtain an optimized cleaning scheme; based on the optimized cleaning scheme, performing multi-stage collaborative cleaning on the chromatograph to obtain cleanliness evaluation characterization; performing graph structure optimization on the cleaning knowledge graph based on the cleanliness evaluation characterization and optimization cleaning scheme to obtain an optimized cleaning knowledge graph; the cleaning precision and efficiency of the chromatographic instrument can be improved.

Description

Technical Field

[0001] This invention relates to the field of data cleaning technology, and in particular to a chromatograph cleaning method, system, smart terminal, and storage medium. Background Technology

[0002] Existing chromatograph cleaning technologies lack comprehensive and in-depth multi-dimensional data acquisition and analysis during the operation of intelligent terminals. They struggle to systematically integrate sensor time-series records, maintenance logs, and historical cleaning case data, resulting in an inability to accurately capture contamination characteristic data and leading to biased judgments about contamination status. Furthermore, cleaning solutions often rely on fixed procedures or experience-based judgments, lacking adaptability to specific contamination situations. This hinders the development of targeted cleaning strategies, resulting in incomplete cleaning or excessive resource consumption.

[0003] Furthermore, existing technologies lack dynamic optimization and knowledge accumulation mechanisms for the cleaning process. After the initial cleaning plan is formulated, there is a lack of multi-scenario simulation verification and potential exploration, making it impossible to avoid unreasonable aspects during execution in advance. Moreover, the feedback on the cleaning results is insufficient to effectively support subsequent plan optimization, leading to insufficient stability in cleaning efficiency and effectiveness. Therefore, improving the accuracy and efficiency of chromatograph cleaning has become an urgent problem to be solved. Summary of the Invention

[0004] This disclosure provides a method, system, intelligent terminal, and storage medium for cleaning a chromatograph.

[0005] In a first aspect, this disclosure provides a method for cleaning a chromatograph, including: S1. Obtain multidimensional operational trace data and historical cleaning cases of the chromatograph, and perform structured analysis on the multidimensional operational trace data to obtain the pollution characteristic data of the chromatograph; S2. Extract the cleaning status features from the historical cleaning cases, perform topology construction on the cleaning status features, and obtain the cleaning knowledge graph of the chromatograph; S3. Based on the pollution characteristic data, perform optimal path deduction on the cleaning knowledge graph to obtain the personalized cleaning trajectory of the chromatograph; S4. Perform decision-making deduction on the personalized cleaning trajectory to obtain the initial cleaning plan for the chromatograph, and perform forward planning on the initial cleaning plan to obtain the optimized cleaning plan for the chromatograph. S5. Based on the optimized cleaning scheme, perform multi-stage synergistic cleaning on the chromatograph to obtain a cleanliness assessment characterization of the chromatograph. S6. Based on the cleanliness assessment and the optimized cleaning scheme, the cleaning knowledge graph is optimized to obtain the optimized cleaning knowledge graph of the chromatograph.

[0006] In a preferred embodiment, the acquisition of multidimensional operational trace data and historical cleaning cases of the chromatograph, and the structured analysis of the multidimensional operational trace data to obtain the contamination characteristic data of the chromatograph, includes: By retrieving the cleaning process records of the chromatograph, historical cleaning cases of the chromatograph can be obtained; The embedded sensor time-series records and maintenance log files of the chromatograph are collected synchronously to obtain multi-dimensional operational trace data of the chromatograph; The multidimensional operational trace data is subjected to feature waveform separation to obtain the pressure fluctuation data and temperature anomaly data of the chromatograph. The spectral characteristic peaks of the multidimensional operation trace data are extracted to obtain the composition data of the inner wall of the chromatograph; The pressure fluctuation characteristic data, the temperature anomaly characteristic data, and the inner wall material composition data are correlated and fused to obtain the contamination characteristic data of the chromatograph.

[0007] In a preferred embodiment, the step of extracting cleaning state features from the historical cleaning cases and performing topology construction on the cleaning state features to obtain the cleaning knowledge graph of the chromatograph includes: Semantic parsing is performed on the historical cleaning cases to obtain a semantic description of the pollution state of the historical cleaning cases; The semantic description of the contamination state is vectorized and encoded to obtain the cleaning state feature vector of the semantic description of the contamination state. Based on the historical cleaning cases, logical deduction is performed on the cleaning state feature vector to obtain the causal transformation relationship of the cleaning state feature vector; Based on the historical cleaning cases, the cleaning state feature vector is subjected to probability fitting to obtain the transition probability of the cleaning state feature vector; Using the cleaning state feature vector as nodes and the causal transformation relationship and the transition probability as edges, a cleaning knowledge graph of the chromatograph is constructed.

[0008] In a preferred embodiment, the step of performing optimal path deduction on the cleaning knowledge graph based on the contamination feature data to obtain the personalized cleaning trajectory of the chromatograph includes: Mapping the pollution feature data to graph nodes yields the target pollution nodes of the pollution feature data; Based on the target contaminated node, path enumeration is performed on the cleaning knowledge graph to obtain candidate cleaning paths for the cleaning knowledge graph. A multi-objective comprehensive evaluation is performed on the candidate cleaning paths to obtain a comprehensive cost index for the candidate cleaning paths; Based on the comprehensive cost index, the candidate cleaning paths are optimized to obtain the personalized cleaning trajectory of the chromatograph.

[0009] In a preferred embodiment, the step of performing decision deduction on the personalized cleaning trajectory to obtain an initial cleaning plan for the chromatograph, and performing forward planning on the initial cleaning plan to obtain an optimized cleaning plan for the chromatograph, includes: Perform time-series logic analysis on the personalized cleaning trajectory to obtain the basic cleaning action sequence of the personalized cleaning trajectory; The basic cleaning action sequence is encapsulated into a strategy to obtain the initial cleaning plan for the chromatograph; The initial cleaning scheme is virtually mapped to obtain a simulated copy of the initial cleaning scheme; The simulation copy is subjected to multi-dimensional parameter perturbation to obtain the virtual execution effect of the initial cleaning scheme under multi-perturbation scenarios; The virtual execution effect is analyzed to obtain the optimization potential of the initial cleaning scheme; The optimization potential is mapped to the initial cleaning scheme to obtain the optimized cleaning scheme for the chromatograph.

[0010] In a preferred embodiment, the step of performing multi-stage synergistic cleaning of the chromatograph based on the optimized cleaning scheme to obtain a cleanliness assessment characterization of the chromatograph includes: Based on the optimized cleaning scheme, the chromatograph is subjected to multi-stage cleaning operations, and the stage cleaning data of the chromatograph is collected. Multidimensional analysis of the stage cleaning data is performed to obtain real-time contaminant load data, real-time pressure and temperature monitoring data, and cumulative cleaning agent consumption data of the stage cleaning data. Based on the current stage data and the previous stage data of the real-time pollutant load data, the current pollutant removal rate increment of the chromatograph is obtained; The current cumulative data of cleaning agent consumption and the real-time monitoring data of pressure and temperature are normalized to obtain the current resource consumption factor of the chromatograph. The state transition efficiency of the chromatograph is determined based on the current pollutant removal rate increment and the causal transformation relationship and transition probability in the cleaning knowledge graph. Based on the historical cleaning cases, a correlation regression analysis was performed on the cleaning state feature vector and the cleaning effect evaluation data of the chromatograph to obtain the dynamic performance weight coefficient of the cleaning state feature vector. The cleanliness assessment characterization of the chromatograph is calculated, wherein the calculation formula for the cleanliness assessment characterization is as follows: The cleanliness assessment characterization of the chromatograph is calculated, wherein the calculation formula for the cleanliness assessment characterization is as follows: ; in, This is a characterization of the cleanliness assessment. This represents the total number of cleaning stages for the chromatograph. The increment of the pollutant removal rate, The state transition efficiency is mentioned above. The current resource consumption factor, For dynamic performance weighting coefficients, This represents the inter-stage coupling gain coefficient. This is the resource consumption penalty coefficient. This is a non-linear amplification factor for resource consumption.

[0011] In a preferred embodiment, the step of optimizing the cleaning knowledge graph based on the cleanliness assessment characterization and the optimized cleaning scheme to obtain the optimized cleaning knowledge graph of the chromatograph includes: The cleanliness assessment characterization was analyzed using performance dimensions to obtain the contribution of the chromatograph. The optimized cleaning scheme is deconstructed and analyzed to obtain the stage characteristics and action parameters of the optimized cleaning scheme; Using the stage features as state nodes in the cleaning knowledge graph and the action parameters as directed edges in the cleaning knowledge graph, the path recording of the chromatograph is obtained. Based on the contribution, the transition probability weights of the directed edges in the path recording are dynamically enhanced, and the attribute vectors of the state nodes in the path recording are collaboratively updated to obtain the optimized cleaning knowledge graph of the chromatograph.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention utilizes a chromatograph cleaning method to perform structured analysis of multidimensional operational trace data, accurately extract contamination characteristic data, construct a cleaning knowledge graph by combining historical cleaning cases, and then generate a personalized cleaning trajectory through optimal path deduction. This allows the cleaning plan to accurately match the actual contamination status of the chromatograph, significantly improving the targeting and effectiveness of cleaning and helping to obtain better cleanliness assessment characteristics.

[0013] 2. By leveraging real-time data acquisition and multi-dimensional analysis during the multi-stage collaborative cleaning process, as well as dynamic optimization based on a knowledge graph of cleanliness assessment and cleaning scheme optimization, this invention can continuously accumulate effective cleaning experience, improve the adaptability and execution efficiency of subsequent cleaning schemes, and optimize resource allocation through forward planning. This ensures reasonable control of resource consumption while guaranteeing cleaning effectiveness, thereby enhancing the overall efficiency of the cleaning process. Attached Figure Description

[0014] The present disclosure will be described in more detail below based on embodiments and with reference to the accompanying drawings: Figure 1 A flowchart illustrating the process of a chromatograph cleaning method according to Embodiment 1 of the present invention is shown. Figure 2 This diagram shows a functional block diagram of a chromatograph cleaning system according to Embodiment 2 of the present invention; Detailed Implementation To enable those skilled in the art to better understand the technical solutions of this disclosure, and to fully understand and implement the process of how this disclosure applies technical means to solve technical problems and achieve corresponding technical effects, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, not all embodiments. The embodiments of this disclosure and the various features within them can be combined with each other without conflict, and the resulting technical solutions are all within the protection scope of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort should fall within the protection scope of this disclosure.

[0015] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0016] Example 1 Figure 1 This is a schematic flowchart illustrating a chromatograph cleaning method provided in an embodiment of this disclosure. Figure 1 As shown, a chromatograph cleaning method includes: S1. Obtain multidimensional operational trace data and historical cleaning cases of the chromatograph, and perform structured analysis on the multidimensional operational trace data to obtain the pollution characteristic data of the chromatograph; In this embodiment of the invention, the acquisition of multidimensional operational trace data and historical cleaning cases of the chromatograph, and the structured analysis of the multidimensional operational trace data to obtain the contamination characteristic data of the chromatograph, includes: By retrieving the cleaning process records of the chromatograph, historical cleaning cases of the chromatograph can be obtained; The embedded sensor time-series records and maintenance log files of the chromatograph are collected synchronously to obtain multi-dimensional operational trace data of the chromatograph; The multidimensional operational trace data is subjected to feature waveform separation to obtain the pressure fluctuation data and temperature anomaly data of the chromatograph. The spectral characteristic peaks of the multidimensional operation trace data are extracted to obtain the composition data of the inner wall of the chromatograph; The pressure fluctuation characteristic data, the temperature anomaly characteristic data, and the inner wall material composition data are correlated and fused to obtain the contamination characteristic data of the chromatograph.

[0017] Review all archived cleaning-related records in the chromatograph's storage system. These records include information such as the time of each cleaning, the corresponding contamination manifestations, the cleaning method used, the type of cleaning agent used, the cleaning duration, and the final cleaning effect feedback. Classify and organize this information according to a unified format to ensure that no key information in each record is omitted, and finally form a complete and standardized collection of historical cleaning cases for the chromatograph.

[0018] Through the built-in data transmission interface of the chromatograph, the various parameter data continuously recorded by the embedded sensor at fixed time intervals during the operation of the smart terminal are retrieved in real time. At the same time, the maintenance log file of the chromatograph is specifically consulted. This file records in detail the daily inspection results, fault repair status, component replacement information and other contents of the smart terminal. The time-series records collected by the sensor and all relevant data in the maintenance log file are synchronized and matched to ensure that the two types of data correspond completely in the time dimension. Then, they are integrated to form multi-dimensional operation trace data covering the operating status parameters and maintenance status of the smart terminal.

[0019] All data related to pressure and temperature are extracted separately from the multidimensional operation trace data. These data are plotted in chronological order to create waveforms showing the changes in values ​​over time. The changes in values ​​in the waveforms are carefully observed to identify the parts of the values ​​that deviate from the standard pressure range during normal operation of the smart terminal. The pressure data corresponding to these deviations is recorded to form pressure fluctuation data. At the same time, based on the plotted waveforms, segments of temperature values ​​that exceed the standard temperature range set by the smart terminal are identified, and the temperature data corresponding to these segments is recorded to form temperature anomaly data.

[0020] Using a professional intelligent spectral analysis terminal, the spectral information contained in the multidimensional operation trace data is processed in a targeted manner. First, the characteristic spectral bands that may correspond to different substances are separated. Then, the changes in the spectral curves in each band are observed one by one, and the characteristic peaks appearing in the curves are identified. Each characteristic peak corresponds to a specific substance component. The type of substance is determined based on the specific position of the characteristic peak, and the relative content of the substance is determined based on the shape and intensity of the characteristic peak. All the identified substance types and their corresponding relative content information are systematically organized to form the inner wall material composition data.

[0021] First, we conduct an in-depth analysis of the correspondence between the amplitude and frequency of pressure fluctuations in the pressure fluctuation characteristic data and the content of various substances in the inner wall material composition data, clarifying the impact of changes in the content of different substances on pressure fluctuations. Then, we explore the intrinsic relationship between the degree and duration of temperature anomalies in the temperature anomaly characteristic data and the inner wall material composition and pressure fluctuations, clarifying the interaction law among the three. Subsequently, we comprehensively integrate all key information related to pollution in these three types of data, eliminate duplicate content, and retain the relevant data that can reflect the nature of pollution, ultimately forming pollution characteristic data that can comprehensively and accurately reflect the pollution status of the chromatograph.

[0022] The beneficial effects are that this implementation process, through standardized historical cleaning case retrieval, comprehensive multi-dimensional operational trace data collection, and precise feature extraction and correlation fusion operations, ensures the integrity, accuracy, and comprehensiveness of pollution feature data. It effectively solves the problem of inaccurate pollution feature capture in existing technologies, and provides a solid and reliable data foundation for subsequent cleaning knowledge graph construction, personalized cleaning trajectory deduction, and optimized cleaning plan formulation, thereby improving the targeting and effectiveness of chromatograph cleaning from the source.

[0023] S2. Extract the cleaning status features from the historical cleaning cases, perform topology construction on the cleaning status features, and obtain the cleaning knowledge graph of the chromatograph; In this embodiment of the invention, the step of extracting cleaning state features from the historical cleaning cases and performing topology construction on the cleaning state features to obtain the cleaning knowledge graph of the chromatograph includes: Semantic parsing is performed on the historical cleaning cases to obtain a semantic description of the pollution state of the historical cleaning cases; The semantic description of the contamination state is vectorized and encoded to obtain the cleaning state feature vector of the semantic description of the contamination state. Based on the historical cleaning cases, logical deduction is performed on the cleaning state feature vector to obtain the causal transformation relationship of the cleaning state feature vector; Based on the historical cleaning cases, the cleaning state feature vector is subjected to probability fitting to obtain the transition probability of the cleaning state feature vector; Using the cleaning state feature vector as nodes and the causal transformation relationship and the transition probability as edges, a cleaning knowledge graph of the chromatograph is constructed.

[0024] By meticulously studying the text of each historical cleaning case, key information related to pollution is identified and extracted, including the specific manifestations of pollution, the location of pollution, the severity of pollution, and the possible causes of pollution. This scattered key information is then organized and integrated according to a unified linguistic logic to form a standardized description that can completely and accurately reflect the pollution status in each case, thus obtaining a semantic description of the pollution status of historical cleaning cases.

[0025] First, identify the core feature dimensions related to chromatograph contamination. These dimensions cover key attributes such as contamination level, contamination location, and contaminant type. Set a fixed descriptive range and corresponding numerical standard for each dimension. Then, in accordance with the semantic description of contamination status, extract the corresponding specific information from each feature dimension and assign corresponding values ​​according to the set numerical standards. Arrange the values ​​corresponding to all dimensions in a preset order to form a numerical sequence that can quantify the semantic description of contamination status, thus obtaining the cleaning status feature vector of the semantic description of contamination status.

[0026] The changes in the cleaning state feature vectors of all historical cleaning cases are collected. The cleaning state feature vectors corresponding to the initial contamination state in each case are analyzed one by one, as well as the cleaning state feature vectors after the transformation after a specific cleaning operation. By comparing the contamination states corresponding to the two feature vectors, it is clear that the contamination state corresponding to the previous feature vector is the cause and the contamination state corresponding to the next feature vector is the effect. Thus, the correlation between different cleaning state feature vectors is determined, that is, the causal transformation relationship of the cleaning state feature vectors is obtained.

[0027] The specific occurrences of causal transformations between feature vectors of the cleanup state in all historical cleanup cases are statistically analyzed. For each pair of feature vectors with a causal transformation relationship, the total number of times this transformation occurs in all cases is recorded. At the same time, the total number of times feature vector X transforms in all cases is also counted. The number of transformations from X to Y is divided by the total number of transformations of X to Y, and the result is the transition probability from feature vector X to feature vector Y. The transition probability between all feature vector pairs with a causal transformation relationship is calculated in this way.

[0028] Each independent cleaning state feature vector is used as a node in the cleaning knowledge graph. Each node is labeled with a corresponding feature vector identifier in the graph. Based on the previously determined causal transformation relationship, two nodes with causal relationship are connected by directed lines. The lines point from the feature vector node representing the "cause" to the feature vector node representing the "effect". At the same time, the corresponding transition probability value is labeled on each directed line. In this way, all nodes and edges are organically integrated to form a complete chromatograph cleaning knowledge graph.

[0029] The beneficial effects are that by analyzing and deeply mining historical cleaning cases layer by layer, the cleaning status characteristics are accurately extracted and a cleaning knowledge graph containing causal transformation relationships and transition probabilities is constructed. This allows historical cleaning experience to be systematically and structurally accumulated, providing comprehensive and reliable knowledge support for subsequent deduction of the optimal cleaning path based on pollution characteristic data. This effectively improves the scientific nature and pertinence of subsequent cleaning plan formulation and avoids subjective bias caused by relying on experience-based judgment.

[0030] S3. Based on the pollution characteristic data, perform optimal path deduction on the cleaning knowledge graph to obtain the personalized cleaning trajectory of the chromatograph; In this embodiment of the invention, the step of performing optimal path deduction on the cleaning knowledge graph based on the pollution feature data to obtain the personalized cleaning trajectory of the chromatograph includes: Mapping the pollution feature data to graph nodes yields the target pollution nodes of the pollution feature data; Based on the target contaminated node, path enumeration is performed on the cleaning knowledge graph to obtain candidate cleaning paths for the cleaning knowledge graph. A multi-objective comprehensive evaluation is performed on the candidate cleaning paths to obtain a comprehensive cost index for the candidate cleaning paths; Based on the comprehensive cost index, the candidate cleaning paths are optimized to obtain the personalized cleaning trajectory of the chromatograph.

[0031] The key information contained in the pollution feature data, such as the degree of pollution, the type of pollutant, and the scope of pollution impact, is compared one by one with the attribute information of all cleaning state feature vector nodes in the cleaning knowledge graph. The focus is on checking the degree of matching of each piece of information, and the cleaning state feature vector node whose attribute information completely matches the pollution feature data is found. This node is the target pollution node corresponding to the pollution feature data.

[0032] Starting from the target contaminated node, following the causal transformation relationships indicated by the directed edges in the cleaning knowledge graph, all possible extended paths are traversed one by one. All nodes passed through each path and the connection relationships between nodes are recorded throughout the process. This ensures that no complete path starting from the target contaminated node and ultimately reaching the node representing the clean state is missed. All these complete paths are collected and organized to form candidate cleaning paths in the cleaning knowledge graph.

[0033] For each candidate cleaning path, an evaluation is conducted from three core dimensions: cleaning effect, resource consumption, and cleaning duration. The cleaning effect is determined by the cleanliness standard corresponding to the end node of the path. Resource consumption is calculated based on the total amount of cleaning agent and energy consumed during the execution of the path. The cleaning duration is calculated based on the total time required for the path from start to finish. The evaluation results of these three dimensions are converted into a uniform evaluation score. Then, according to a preset fixed importance ratio, the scores of each dimension are added together to obtain the comprehensive cost index of each candidate cleaning path.

[0034] By comparing the comprehensive cost index of all candidate cleaning paths horizontally and checking the index value of each path one by one, the candidate cleaning path with the smallest comprehensive cost index value is selected. This path can achieve the optimal balance between resource consumption and time cost while ensuring that the cleaning effect meets the standard. This path is the personalized cleaning trajectory of the chromatograph.

[0035] The beneficial effects are that the accurate mapping of graph nodes enables effective connection between pollution characteristics and knowledge graphs, comprehensive path enumeration ensures that no potential optimal solutions are missed, multi-dimensional comprehensive evaluation makes the judgment of path merit more scientific, and the final path optimization accurately locks in personalized trajectories that are suitable for the current pollution status. This effectively avoids the blindness of traditional fixed-path cleaning, greatly improves the targeting and rationality of cleaning paths, and lays a solid foundation for the optimization of subsequent cleaning schemes.

[0036] S4. Perform decision-making deduction on the personalized cleaning trajectory to obtain the initial cleaning plan for the chromatograph, and perform forward planning on the initial cleaning plan to obtain the optimized cleaning plan for the chromatograph. In this embodiment of the invention, the step of performing decision deduction on the personalized cleaning trajectory to obtain an initial cleaning plan for the chromatograph, and performing forward planning on the initial cleaning plan to obtain an optimized cleaning plan for the chromatograph, includes: Perform time-series logic analysis on the personalized cleaning trajectory to obtain the basic cleaning action sequence of the personalized cleaning trajectory; The basic cleaning action sequence is encapsulated into a strategy to obtain the initial cleaning plan for the chromatograph; The initial cleaning scheme is virtually mapped to obtain a simulated copy of the initial cleaning scheme; The simulation copy is subjected to multi-dimensional parameter perturbation to obtain the virtual execution effect of the initial cleaning scheme under multi-perturbation scenarios; The virtual execution effect is analyzed to obtain the optimization potential of the initial cleaning scheme; The optimization potential is mapped to the initial cleaning scheme to obtain the optimized cleaning scheme for the chromatograph.

[0037] The personalized cleaning trajectory is organized according to the logic of time progression, clarifying the sequence and logical relationship of each link in the trajectory. The specific cleaning operations corresponding to the trajectory are broken down one by one. These operations include the injection method of cleaning agent, the rinsing path inside the smart terminal, the duration of each cleaning link, and the control actions of pressure and temperature. These specific cleaning operations are arranged in the original time sequence to form a continuous basic cleaning action sequence with clear execution logic.

[0038] The basic cleaning action sequence is used as the core content, supplemented with the execution conditions of each cleaning action, such as the pressure range and temperature standards that the intelligent terminal must meet when the action is started. The connection requirements between actions are clarified, that is, the judgment criteria for the completion of the previous action and the triggering conditions for the start of the next action. At the same time, the safety specifications that must be followed during the cleaning process are integrated, such as the precautions for the use of cleaning agents and the protection requirements for the operation of intelligent terminals. These contents are organically combined with the basic cleaning action sequence to form a chromatograph initial cleaning plan that is structurally complete, logically clear and directly executable.

[0039] A virtual simulation platform consistent with the actual chromatograph operating environment is built. Each action, execution condition, and connection logic in the initial cleaning plan is transformed into a digital model element that the virtual simulation platform can recognize and run. This accurately restores the complete execution process of the initial cleaning plan, including the timing relationship of actions, the control logic of parameters, and the rules for changes in the state of the intelligent terminal. Finally, a simulation copy that completely corresponds to the initial cleaning plan is formed in the virtual simulation platform.

[0040] Key parameters affecting cleaning effectiveness and resource consumption are identified, including cleaning agent concentration, cleaning temperature, cleaning pressure, and action execution time. Within a reasonable range for each parameter, the specific values ​​of the parameters are adjusted sequentially to create multiple perturbation scenarios with different parameter combinations. Simulation copies of each perturbation scenario are run in a virtual simulation platform, and cleaning effectiveness data such as the proportion of contaminant removal and the cleanliness of smart terminals are recorded in real time, as well as resource consumption data such as cleaning agent dosage and energy consumption. The results of these records constitute the virtual execution effect of the initial cleaning plan under multiple perturbation scenarios.

[0041] The virtual execution effect data under all multi-perturbation scenarios are classified and organized. The differences in cleaning effect and resource consumption under different parameter combinations are compared to identify typical scenarios such as optimal cleaning effect but high resource consumption, satisfactory cleaning effect and moderate resource consumption, and poor cleaning effect but low resource consumption. The influence of parameter adjustment on execution effect in these scenarios is analyzed to clarify which parameters in the initial cleaning scheme have room for adjustment and which action execution logic can be optimized. In this way, the optimization potential of the initial cleaning scheme in terms of improving cleaning effect and reducing resource consumption is determined.

[0042] Based on the identified optimization potential, specific modifications were made to the adjustable parameters and execution logic in the initial cleaning scheme. For example, regarding the cleaning agent concentration parameter, if it was found that appropriately increasing the concentration could significantly improve the cleaning effect without excessively increasing resource consumption, the setting value of this parameter was adjusted. Regarding the action execution logic, if it was found that there was time redundancy in the connection between two actions, the connection triggering conditions were optimized to shorten the interval time. All optimization adjustments were implemented one by one into the initial cleaning scheme to form an optimized chromatograph cleaning scheme that takes into account both cleaning effect and resource utilization efficiency.

[0043] The beneficial effects are that the abstract cleaning trajectory is transformed into a specific and executable sequence of actions through temporal logic parsing, and a standardized initial cleaning scheme is formed through strategy encapsulation. Then, the initial scheme is simulated and verified in multiple scenarios with the help of virtual mapping and multi-dimensional parameter perturbation. The optimization potential is accurately explored and the scheme is iterated. This effectively avoids unreasonable problems that may occur in the actual execution of the initial scheme. The final optimized cleaning scheme not only fits the individual contamination status of the chromatograph, but also achieves the optimal balance between cleaning effect and resource consumption, which greatly improves the scientificity and feasibility of the cleaning scheme.

[0044] S5. Based on the optimized cleaning scheme, perform multi-stage synergistic cleaning on the chromatograph to obtain a cleanliness assessment characterization of the chromatograph. In this embodiment of the invention, the step of performing multi-stage synergistic cleaning of the chromatograph based on the optimized cleaning scheme to obtain a cleanliness assessment characterization of the chromatograph includes: Based on the optimized cleaning scheme, the chromatograph is subjected to multi-stage cleaning operations, and the stage cleaning data of the chromatograph is collected. Multidimensional analysis of the stage cleaning data is performed to obtain real-time contaminant load data, real-time pressure and temperature monitoring data, and cumulative cleaning agent consumption data of the stage cleaning data. Based on the current stage data and the previous stage data of the real-time pollutant load data, the current pollutant removal rate increment of the chromatograph is obtained; The current cumulative data of cleaning agent consumption and the real-time monitoring data of pressure and temperature are normalized to obtain the current resource consumption factor of the chromatograph. The state transition efficiency of the chromatograph is determined based on the current pollutant removal rate increment and the causal transformation relationship and transition probability in the cleaning knowledge graph. Based on the historical cleaning cases, a correlation regression analysis was performed on the cleaning state feature vector and the cleaning effect evaluation data of the chromatograph to obtain the dynamic performance weight coefficient of the cleaning state feature vector. The cleanliness assessment characterization of the chromatograph is calculated, wherein the calculation formula for the cleanliness assessment characterization is as follows: The cleanliness assessment characterization of the chromatograph is calculated, wherein the calculation formula for the cleanliness assessment characterization is as follows: ; in, This is a characterization of the cleanliness assessment. This represents the total number of cleaning stages for the chromatograph. The increment of the pollutant removal rate, The state transition efficiency is mentioned above. The current resource consumption factor, For dynamic performance weighting coefficients, This represents the inter-stage coupling gain coefficient. This is the resource consumption penalty coefficient. This is a non-linear amplification factor for resource consumption.

[0045] Following the cleaning phase sequence defined in the optimized cleaning plan, each phase of the cleaning operation is performed sequentially. Each phase strictly adheres to the cleaning agent type, dosage, cleaning time, pressure, and temperature control standards specified in the plan. During and after each phase of the cleaning operation, various data related to that phase of the cleaning are comprehensively collected through a combination of the intelligent terminal's built-in monitoring device and manual recording. This includes changes in contaminants, intelligent terminal operating parameters, and cleaning agent consumption, forming complete chromatograph phase cleaning data.

[0046] The collected stage cleaning data is classified and broken down. Data that reflects the total amount of pollutants in the current smart terminal is selected and organized to form real-time pollutant load data. Continuous records of internal pressure and temperature of the smart terminal during the cleaning process are extracted and arranged in chronological order to form real-time pressure and temperature monitoring data. The total amount of cleaning agent consumed from the beginning to the end of each stage is calculated and summarized to form cumulative cleaning agent consumption data.

[0047] Extract the total pollutant load data recorded at the end of the current stage from the real-time pollutant load data, subtract the total pollutant load data recorded at the end of the previous stage, and obtain the amount of pollutant reduction in the current stage compared to the previous stage. Divide this reduction in pollutant load by the total pollutant load at the end of the previous stage; the result is the current pollutant removal rate increment of the chromatograph. First, define the standard reference ranges for cleaning agent consumption, pressure, and temperature. Compare the cumulative cleaning agent consumption data of the current stage with the maximum standard reference value of this data to calculate the relative proportion of cleaning agent consumption. Similarly, take the average value of the real-time pressure monitoring data of the current stage and compare it with the maximum value of the pressure standard reference range to obtain the relative pressure proportion. Take the average value of the real-time temperature monitoring data of the current stage and compare it with the maximum value of the temperature standard reference range to obtain the relative temperature proportion. Add these three relative proportions in the same proportion; the sum is the current resource consumption factor of the chromatograph.

[0048] In the cleaning knowledge graph, the starting node corresponding to the current pollution state and the ending node corresponding to the target state after cleaning are found. The causal transformation relationship between these two nodes and the corresponding transition probability are determined. The current pollutant removal rate increment is compared with the expected removal effect corresponding to the transition probability. If the current pollutant removal rate increment reaches the expected effect, the state transition efficiency is consistent with the transition probability. If the current pollutant removal rate increment is higher than the expected effect, the state transition efficiency is higher than the transition probability. If it is lower than the expected effect, the state transition efficiency is lower than the transition probability. Based on this, the state transition efficiency of the current stage is determined.

[0049] All historical cleaning cases were collected, and the cleaning status feature vectors and corresponding cleaning effect evaluation data of each case were extracted. The influence of each cleaning status feature vector on the cleaning effect in different cases was analyzed. If a certain cleaning status feature vector can significantly affect the cleaning effect in multiple cases and the corresponding cleaning effect evaluation data is generally high, it indicates that the feature vector is more important and is assigned a higher weight coefficient. Conversely, a lower weight coefficient is assigned. Through this correlation analysis, the dynamic effectiveness weight coefficient of each cleaning status feature vector is determined.

[0050] This calculation comprehensively reflects the cleanliness of the chromatograph after multi-stage collaborative cleaning. It takes into account the pollutant removal effect, state transition efficiency, resource consumption, and dynamic performance weight of each stage in the multi-stage cleaning process, providing a key basis for the subsequent optimization of the cleaning knowledge graph.

[0051] The total number of cleaning stages is determined when performing multi-stage collaborative cleaning of the chromatograph. The increase in contaminant removal rate is obtained by comparing the current stage data with the previous stage data using real-time contaminant load data, which is derived from multi-dimensional analysis of the stage cleaning data collected during the multi-stage cleaning operation. The state transition efficiency is determined based on the increase in contaminant removal rate and the causal transformation relationship and transition probability in the cleaning knowledge graph. The causal transformation relationship is obtained through logical deduction of the cleaning state feature vectors from historical cleaning cases, and the transition probability is obtained through probability fitting of the cleaning state feature vectors. The cleaning state feature vectors are obtained by vectorizing the semantic descriptions of the contamination states in historical cleaning cases. The current resource consumption factor is obtained by normalizing the cumulative data of cleaning agent consumption in the current stage and the real-time monitoring data of pressure and temperature; these data are all derived from multi-dimensional analysis of the stage cleaning data. The dynamic efficiency weighting coefficient is obtained by correlating the cleaning state feature vectors with the chromatograph's cleaning effect evaluation data based on historical cleaning cases, finding the relationship between the two through regression analysis, and determining the coefficient that minimizes the error between them. The inter-stage coupling gain coefficient, resource consumption penalty coefficient, and nonlinear amplification factor of resource consumption are fixed coefficients set in the formula to adjust the degree of influence of relevant factors on the cleanliness assessment characterization.

[0052] When the increase in contaminant removal rate increases, the cumulative sum of the numerator increases, and the cleanliness assessment characterization rises accordingly. Improved state transition efficiency increases the corresponding term in the numerator, further enhancing the cleanliness assessment characterization. Increasing the dynamic efficiency weighting coefficient increases the contribution of the corresponding stage in the numerator; if the contaminant removal rate increment and state transition efficiency of that stage are at a reasonable level, the cleanliness assessment characterization will rise. An increase in the inter-stage coupling gain coefficient amplifies the positive impact of state transition efficiency on the numerator, leading to a rise in the cleanliness assessment characterization. When the cumulative sum of the current resource consumption factor increases, the denominator increases, and the cleanliness assessment characterization decreases accordingly. Increases in the resource consumption penalty coefficient and the nonlinear amplification factor of resource consumption further amplify the impact of resource consumption on the denominator, resulting in a decrease in the cleanliness assessment characterization.

[0053] The beneficial effects are that this implementation process, through standardized execution of multi-stage collaborative cleaning and comprehensive data collection, combined with detailed multi-dimensional analysis and scientific parameter calculation, can accurately and comprehensively obtain various key data reflecting the cleaning effect of the chromatograph. The resulting cleanliness assessment characterization can truly and objectively reflect the actual cleanliness status of the chromatograph, providing a reliable basis for the subsequent optimization of the cleaning knowledge graph. At the same time, it makes the evaluation of the cleaning effect more quantitative and accurate, avoids the bias caused by subjective judgment, and improves the scientificity and traceability of the entire cleaning process.

[0054] S6. Based on the cleanliness assessment and the optimized cleaning scheme, the cleaning knowledge graph is optimized to obtain the optimized cleaning knowledge graph of the chromatograph.

[0055] In this embodiment of the invention, the step of optimizing the cleaning knowledge graph structure based on the cleanliness assessment characterization and the optimized cleaning scheme to obtain the optimized cleaning knowledge graph of the chromatograph includes: The cleanliness assessment characterization was analyzed using performance dimensions to obtain the contribution of the chromatograph. The optimized cleaning scheme is deconstructed and analyzed to obtain the stage characteristics and action parameters of the optimized cleaning scheme; Using the stage features as state nodes in the cleaning knowledge graph and the action parameters as directed edges in the cleaning knowledge graph, the path recording of the chromatograph is obtained. Based on the contribution, the transition probability weights of the directed edges in the path recording are dynamically enhanced, and the attribute vectors of the state nodes in the path recording are collaboratively updated to obtain the optimized cleaning knowledge graph of the chromatograph.

[0056] Starting from three performance dimensions—improved cleaning effect, controlled resource consumption, and optimized cleaning efficiency—this paper analyzes the actual supporting role of the cleanliness assessment characterization in each dimension against the specific values ​​of the characterization. It determines whether the cleanliness assessment characterization in each dimension meets or exceeds the expected standard. Based on the analysis results of the three dimensions, the paper determines the positive effect of the cleanliness assessment characterization on the improvement of the chromatograph cleaning effect. This positive effect is the contribution of the chromatograph.

[0057] The optimized cleaning plan is broken down into predetermined cleaning stages. The key points of pollution treatment, the target state to be achieved, and the connection requirements between the current stage and the previous and subsequent stages constitute the stage characteristics of the optimized cleaning plan. At the same time, the specific operational details involved in the execution of each stage are carefully extracted, including the specific amount of cleaning agent used, the execution time of the cleaning action, and the pressure and temperature control standards of the intelligent terminal. These specific operational details are the action parameters of the optimized cleaning plan.

[0058] Each stage feature obtained from the decomposition is compared with the attribute information of all state nodes in the cleaning knowledge graph. If the pollution state or target state corresponding to a certain stage feature is completely consistent with the attribute of an existing state node, the stage feature is directly associated with that state node. If there is no matching existing node, a new state node is created based on the stage feature, thus completing the mapping from stage features to knowledge graph state nodes. The action parameters of each stage are mapped to the directed edges connecting the current stage state node and the next stage state node, and the specific action parameters corresponding to each directed edge are defined. In this way, a complete chromatograph path recording of the cleaning path and corresponding parameters is formed.

[0059] The adjustment range is determined based on the specific value of the contribution. The higher the contribution, the better the corresponding path recording is in actual cleaning. The transition probability weight of each directed edge in the path recording is increased according to this adjustment range, making it easier to prioritize the efficient path in subsequent path deduction. At the same time, combined with the actual performance of the stage characteristics during the execution of the optimized cleaning scheme, the feature values ​​related to the pollution type, pollution degree, and treatment difficulty in the attribute vector of the corresponding state node are adjusted so that the attribute vector of the state node is more in line with the needs of the actual cleaning scenario. The path recording after the transition probability weight enhancement and state node attribute vector update is integrated into the original cleaning knowledge graph to form the optimized cleaning knowledge graph of the chromatograph.

[0060] The beneficial effects are that the actual value of cleanliness assessment is accurately quantified through the efficiency dimension, the core information of the optimized cleaning plan is fully extracted through deconstruction analysis, the plan and knowledge graph are effectively connected through path recording, and the dynamic optimization based on contribution makes the node attributes and edge transition probabilities of the knowledge graph more in line with the actual needs of efficient cleaning. This allows the optimized cleaning knowledge graph to continuously accumulate high-quality cleaning experience, significantly improve the accuracy and efficiency of subsequent cleaning path deduction and plan formulation, and form a virtuous cycle of cleaning knowledge iteration.

[0061] Example 2 like Figure 2 As shown in the figure, this embodiment also provides a functional block diagram of a chromatograph cleaning system.

[0062] The chromatograph cleaning system 100 described in this embodiment can be installed in a smart terminal. Depending on the functions implemented, the chromatograph cleaning system 100 may include a contamination feature data identification module 101, a cleaning knowledge graph construction module 102, a cleaning path deduction module 103, a cleaning plan generation module 104, a collaborative execution module 105, and a graph structure optimization module 106. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the smart terminal processor and perform fixed functions, stored in the smart terminal's memory.

[0063] In this embodiment, the functions of each module / unit are as follows: The pollution feature data identification module 101 is used to acquire multidimensional operation trace data and historical cleaning cases of the chromatograph, and to perform structured analysis on the multidimensional operation trace data to obtain the pollution feature data of the chromatograph. The cleaning knowledge graph construction module 102 is used to extract cleaning status features from the historical cleaning cases, perform topology construction on the cleaning status features, and obtain the cleaning knowledge graph of the chromatograph. The cleaning path deduction module 103 is used to perform optimal path deduction on the cleaning knowledge graph based on the pollution feature data to obtain the personalized cleaning trajectory of the chromatograph. The cleaning scheme generation module 104 is used to perform decision-making and deduction on the personalized cleaning trajectory to obtain the initial cleaning scheme of the chromatograph, and to perform forward planning on the initial cleaning scheme to obtain the optimized cleaning scheme of the chromatograph. The collaborative execution module 105 is used to perform multi-stage collaborative cleaning of the chromatograph based on the cleaning optimization trajectory, so as to obtain a cleanliness assessment characterization of the chromatograph. The spectral structure optimization module 106 is used to optimize the spectral structure of the cleaning knowledge graph based on the cleanliness assessment characterization and the cleaning optimization trajectory, so as to obtain the optimized cleaning knowledge graph of the chromatograph.

[0064] In detail, each module of the chromatograph cleaning system 100 described in the embodiments of the present invention uses the same technical means as the chromatograph cleaning method described in Embodiment 1 and Embodiment 2, and can produce the same technical effect, which will not be repeated here.

[0065] In the several embodiments provided by this invention, it should be understood that the disclosed smart terminals, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0066] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0067] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0068] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0069] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0070] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for cleaning a chromatograph, characterized in that, The method includes: S1. Obtain multidimensional operational trace data and historical cleaning cases of the chromatograph, and perform structured analysis on the multidimensional operational trace data to obtain the pollution characteristic data of the chromatograph; S2. Extract the cleaning status features from the historical cleaning cases, perform topology construction on the cleaning status features, and obtain the cleaning knowledge graph of the chromatograph; S3. Based on the pollution characteristic data, perform optimal path deduction on the cleaning knowledge graph to obtain the personalized cleaning trajectory of the chromatograph; S4. Perform decision-making deduction on the personalized cleaning trajectory to obtain the initial cleaning plan for the chromatograph, and perform forward planning on the initial cleaning plan to obtain the optimized cleaning plan for the chromatograph. S5. Based on the optimized cleaning scheme, perform multi-stage synergistic cleaning on the chromatograph to obtain a cleanliness assessment characterization of the chromatograph. S6. Based on the cleanliness assessment and the optimized cleaning scheme, the cleaning knowledge graph is optimized to obtain the optimized cleaning knowledge graph of the chromatograph.

2. The chromatograph cleaning method as described in claim 1, characterized in that, The process involves acquiring multidimensional operational trace data and historical cleaning cases of the chromatograph, and performing structured analysis on the multidimensional operational trace data to obtain the contamination characteristic data of the chromatograph, including: By retrieving the cleaning process records of the chromatograph, historical cleaning cases of the chromatograph can be obtained; The embedded sensor time-series records and maintenance log files of the chromatograph are collected synchronously to obtain multi-dimensional operational trace data of the chromatograph; The multidimensional operational trace data is subjected to feature waveform separation to obtain the pressure fluctuation data and temperature anomaly data of the chromatograph. The spectral characteristic peaks of the multidimensional operation trace data are extracted to obtain the composition data of the inner wall of the chromatograph; The pressure fluctuation characteristic data, the temperature anomaly characteristic data, and the inner wall material composition data are correlated and fused to obtain the contamination characteristic data of the chromatograph.

3. The chromatograph cleaning method as described in claim 1, characterized in that, The step of extracting cleaning status features from the historical cleaning cases and performing topology construction on the cleaning status features to obtain the cleaning knowledge graph of the chromatograph includes: Semantic parsing is performed on the historical cleaning cases to obtain a semantic description of the pollution state of the historical cleaning cases; The semantic description of the contamination state is vectorized and encoded to obtain the cleaning state feature vector of the semantic description of the contamination state. Based on the historical cleaning cases, logical deduction is performed on the cleaning state feature vector to obtain the causal transformation relationship of the cleaning state feature vector; Based on the historical cleaning cases, the cleaning state feature vector is subjected to probability fitting to obtain the transition probability of the cleaning state feature vector; Using the cleaning state feature vector as nodes and the causal transformation relationship and the transition probability as edges, a cleaning knowledge graph of the chromatograph is constructed.

4. The chromatograph cleaning method as described in claim 1, characterized in that, The step of performing optimal path deduction on the cleaning knowledge graph based on the pollution characteristic data to obtain the personalized cleaning trajectory of the chromatograph includes: Mapping the pollution feature data to graph nodes yields the target pollution nodes of the pollution feature data; Based on the target contaminated node, path enumeration is performed on the cleaning knowledge graph to obtain candidate cleaning paths for the cleaning knowledge graph. A multi-objective comprehensive evaluation is performed on the candidate cleaning paths to obtain a comprehensive cost index for the candidate cleaning paths; Based on the comprehensive cost index, the candidate cleaning paths are optimized to obtain the personalized cleaning trajectory of the chromatograph.

5. The chromatograph cleaning method as described in claim 1, characterized in that, The process of performing decision-making deduction on the personalized cleaning trajectory to obtain an initial cleaning plan for the chromatograph, and then performing forward planning on the initial cleaning plan to obtain an optimized cleaning plan for the chromatograph, includes: Perform time-series logic analysis on the personalized cleaning trajectory to obtain the basic cleaning action sequence of the personalized cleaning trajectory; The basic cleaning action sequence is encapsulated into a strategy to obtain the initial cleaning plan for the chromatograph; The initial cleaning scheme is virtually mapped to obtain a simulated copy of the initial cleaning scheme; The simulation copy is subjected to multi-dimensional parameter perturbation to obtain the virtual execution effect of the initial cleaning scheme under multi-perturbation scenarios; The virtual execution effect is analyzed to obtain the optimization potential of the initial cleaning scheme; The optimization potential is mapped to the initial cleaning scheme to obtain the optimized cleaning scheme for the chromatograph.

6. The chromatograph cleaning method as described in claim 1, characterized in that, The process of performing multi-stage synergistic cleaning on the chromatograph based on the optimized cleaning scheme to obtain a cleanliness assessment characterization of the chromatograph includes: Based on the optimized cleaning scheme, the chromatograph is subjected to multi-stage cleaning operations, and the stage cleaning data of the chromatograph is collected. Multidimensional analysis of the stage cleaning data is performed to obtain real-time contaminant load data, real-time pressure and temperature monitoring data, and cumulative cleaning agent consumption data of the stage cleaning data. Based on the current stage data and the previous stage data of the real-time pollutant load data, the current pollutant removal rate increment of the chromatograph is obtained; The current cumulative data of cleaning agent consumption and the real-time monitoring data of pressure and temperature are normalized to obtain the current resource consumption factor of the chromatograph. The state transition efficiency of the chromatograph is determined based on the current pollutant removal rate increment and the causal transformation relationship and transition probability in the cleaning knowledge graph. Based on the historical cleaning cases, a correlation regression analysis was performed on the cleaning state feature vector and the cleaning effect evaluation data of the chromatograph to obtain the dynamic performance weight coefficient of the cleaning state feature vector. The cleanliness assessment characterization of the chromatograph is calculated, wherein the calculation formula for the cleanliness assessment characterization is as follows: The cleanliness assessment characterization of the chromatograph is calculated, wherein the calculation formula for the cleanliness assessment characterization is as follows: ； in, This is a characterization of the cleanliness assessment. This represents the total number of cleaning stages for the chromatograph. The increment of the pollutant removal rate, The state transition efficiency is mentioned above. The current resource consumption factor, For dynamic performance weighting coefficients, This represents the inter-stage coupling gain coefficient. This is the resource consumption penalty coefficient. This is a non-linear amplification factor for resource consumption.

7. The chromatograph cleaning method as described in claim 1, characterized in that, The process of optimizing the cleaning knowledge graph based on the cleanliness assessment and the optimized cleaning scheme to obtain an optimized cleaning knowledge graph for the chromatograph includes: The cleanliness assessment characterization was analyzed using performance dimensions to obtain the contribution of the chromatograph. The optimized cleaning scheme is deconstructed and analyzed to obtain the stage characteristics and action parameters of the optimized cleaning scheme; Using the stage features as state nodes in the cleaning knowledge graph and the action parameters as directed edges in the cleaning knowledge graph, the path recording of the chromatograph is obtained. Based on the contribution, the transition probability weights of the directed edges in the path recording are dynamically enhanced, and the attribute vectors of the state nodes in the path recording are collaboratively updated to obtain the optimized cleaning knowledge graph of the chromatograph.

8. A chromatograph cleaning system, characterized in that, The system for implementing the chromatograph cleaning method according to claim 1 includes: The pollution characteristic data identification module is used to acquire multidimensional operating trace data and historical cleaning cases of the chromatograph, and to perform structured analysis on the multidimensional operating trace data to obtain the pollution characteristic data of the chromatograph. The cleaning knowledge graph construction module is used to extract cleaning status features from the historical cleaning cases, perform topology construction on the cleaning status features, and obtain the cleaning knowledge graph of the chromatograph. The cleaning path deduction module is used to perform optimal path deduction on the cleaning knowledge graph based on the pollution feature data to obtain the personalized cleaning trajectory of the chromatograph. The cleaning plan generation module is used to perform decision-making and deduction on the personalized cleaning trajectory to obtain the initial cleaning plan of the chromatograph, and to perform forward planning on the initial cleaning plan to obtain the optimized cleaning plan of the chromatograph. The collaborative execution module is used to perform multi-stage collaborative cleaning of the chromatograph based on the cleaning optimization trajectory, so as to obtain a cleanliness assessment characterization of the chromatograph. The spectral structure optimization module is used to optimize the spectral structure of the cleaning knowledge graph based on the cleanliness assessment characterization and the cleaning optimization trajectory, so as to obtain the optimized cleaning knowledge graph of the chromatograph.

9. A computer intelligent terminal, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 7.

Images