Refining heating furnace risk supervision method and system

Abstract

The embodiment of the invention provides a refining heating furnace risk supervision method and system, and belongs to the technical field of equipment risk supervision. The risk supervision method for the refining and chemical heating furnace is characterized by comprising the following steps: collecting real-time operation data of the refining and chemical heating furnace, and preprocessing the real-time operation data to obtain preprocessed data; classifying the preprocessed data based on risk types to obtain a data set under each risk type; determining risk levels of the corresponding risk types one by one based on the data sets under the risk types, and calling corresponding supervision schemes in sequence for the determined risk levels based on a preset risk knowledge graph; and executing the called supervision scheme, and outputting corresponding feedback information based on an execution result.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment risk supervision, and in particular to a refining heating furnace risk supervision method and a refining heating furnace risk supervision system. Background Art

[0002] The refining heating furnace is one of the core equipment of the refining enterprise, and its operating stability is directly related to the operating stability of the entire refining enterprise. In order to ensure the effective progress of the refining process, it is essential to continuously monitor the operating stability of the refining heating furnace.

[0003] In the existing schemes, the most common supervision scheme for refining heating furnaces is to analyze the operation data, identify the faulty equipment, inspect and troubleshoot the faulty equipment, and restart the operation after troubleshooting. This method often requires a large number of shutdowns for maintenance. During the maintenance process, the cause of the fault is reversely fuzzy located based on the abnormal data, and then the fault is troubleshooted based on the fuzzy positioning results to solve the fault problem. But in fact, in the actual operation process, although some components are abnormal, they have a certain ability to continue working, such as simple equipment aging. The existing schemes often qualitatively analyze whether there is a fault in the equipment, and they cannot quantitatively evaluate the risk status of the equipment. For the schemes that can still work continuously, they can only directly replace or continue to operate based on experience. As for when to replace in order to maximize the benefits of cost control, the existing schemes are obviously unable to meet the requirements. In view of the problem that the existing risk supervision schemes for refining heating furnaces cannot quantitatively supervise risks, it is necessary to create a new risk supervision scheme for refining heating furnaces. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a method and system for risk supervision of a refining and heating furnace, so as to at least solve the problem that the existing risk supervision schemes for refining and heating furnaces cannot quantitatively supervise the risks.

[0005] In order to achieve the above-mentioned objectives, the first aspect of the present invention provides a risk supervision method for a refining and heating furnace, the method comprising: collecting real-time operation data of the refining and heating furnace, and preprocessing the real-time operation data to obtain preprocessed data; classifying the preprocessed data based on risk type to obtain data sets under each risk type; determining the risk level of the corresponding risk type based on the data sets under each risk type one by one, and based on a preset risk knowledge graph, calling corresponding supervision plans in sequence for the determined risk levels; executing the called supervision plan, and outputting corresponding feedback information based on the execution result.

[0006] Optionally, the real-time operation data is preprocessed, including: filtering the real-time operation data based on the component type of the refining heating furnace, retaining valid data as the first data set; parsing the data in the first data set to unify the format of all data, and using the parsed data as the second data set; dedimensionalizing the data in the second data set to obtain evaluation data, and using the evaluation data as the third data set.

[0007] Optionally, the real-time operation data is filtered based on the component type of the refining heating furnace, including: for each component type, extracting a data topology tree corresponding to the component type; based on the leaf nodes in the data topology tree, extracting valid data from the real-time operation data one by one; traversing all component types and all real-time operation data, obtaining all valid data, and filtering the remaining data.

[0008] Optionally, the method also includes: constructing a data topology tree for each component type, including: using the operating data directly related to each component type as the parent node data; using the related data that directly affects the parent node data as the child node data of the corresponding parent node, thereby extending the nodes to form an initial topology tree; in the initial topology tree, identifying the in-degree of all nodes, and deleting other branches after retaining any branch for nodes with an in-degree greater than 1 to obtain a second topology tree; in the second topology tree, identifying the data of all nodes to determine whether there are nodes with the same data. If so, retain the node with the highest level and delete other nodes with the same data as the node to obtain a data topology tree for each component type.

[0009] Optionally, the preprocessed data is classified based on the risk type, including: classifying the preprocessed data based on the component types involved in each risk type to obtain preprocessed data of all component types under each risk type; sorting the preprocessed data under each risk type based on a preset level of each risk type, the higher the preset level of the risk type, the higher the corresponding data is sorted; and obtaining a data set under each risk type based on the sorted classification results.

[0010] Optionally, the risk level of the corresponding risk type is determined based on the data set under each risk type one by one, including: simulating the corresponding risk type data based on the data set under each risk type to obtain simulation results; comparing the simulation results of each risk type with a pre-built normal refining and heating furnace simulation model to obtain a deviation value; performing a weighted operation based on a preset level of each risk type and the deviation value to obtain a priority evaluation score corresponding to each risk type; wherein the weight of the preset level of each risk type and the weight corresponding to the deviation value are both obtained based on a preset weight table; comparing the priority evaluation scores of each risk type, the higher the score, the higher the corresponding risk level.

[0011] Optionally, based on the preset risk knowledge graph, calling the corresponding regulatory plan for the determined risk level includes: starting from the risk type with the highest risk level, locating the risk among the risk types, and obtaining risk component information; searching in the preset risk knowledge graph based on the risk component information to obtain the regulatory plan corresponding to the current risk component; traversing all risk types based on the risk level sorting to obtain the regulatory plans corresponding to all risk components.

[0012] Optionally, the risk types include any multiple of furnace tube risk type, furnace nozzle risk type, furnace structure risk type and monitoring equipment risk type.

[0013] Optionally, the supervision plan for executing the call includes: for the located risk components, while ensuring that the same regular monitoring data is collected as other components of the same type, redundant collection of corresponding risk component monitoring data is performed; wherein the redundant collection includes increasing the sampling frequency or extending the sampling period; using the regular collection data and the redundant collection data together as the monitoring data of the risk components, and performing real-time monitoring of the operating status of the risk components based on the monitoring data.

[0014] Optionally, the corresponding feedback information is output based on the execution result, including: if the real-time monitoring result of the risk component shows that it is continuously in a risky state, then a monitoring data report is continuously generated, and the corresponding data report is fed back to the regulatory end once every predetermined period; if the real-time monitoring result of the risk component shows that it is out of the risky state, then the regulatory plan for the risk component is stopped, and a corresponding risk-free signal is output to the regulatory end; if the real-time monitoring result of the risk component shows that its risk level has increased, then the alarm information is continuously output, and the alarm information is pushed to the regulatory end.

[0015] The second aspect of the present invention provides a risk supervision system for a refining and heating furnace, the system comprising: a collection unit, used to collect real-time operating data of the refining and heating furnace, and preprocess the real-time operating data to obtain preprocessed data; a processing unit, used to classify the preprocessed data based on the risk type, and obtain a data set under each risk type; a scheme calling unit, used to determine the risk level of the corresponding risk type based on the data set under each risk type one by one, and based on a preset risk knowledge graph, sequentially call the corresponding supervision scheme for the determined risk level; an execution unit, used to execute the called supervision scheme, and output corresponding feedback information based on the execution result.

[0016] On the other hand, the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed on a computer, enables the computer to execute the above-mentioned refining heating furnace risk supervision method.

[0017] Through the above technical scheme, the scheme of the present invention first comprehensively collects the operation data of the refining heating furnace, and then performs data preprocessing, retains valid data, and reduces the data volume. Then, data classification is performed based on the risk type to facilitate subsequent differentiated risk analysis. The risk level of the risk type is determined based on the specific data analysis results, and the supervision scheme is called based on the risk level to achieve differentiated supervision of the risk level and quantitative differentiation of risks. Then, risks are continuously supervised based on the supervision scheme, and the supervision results are fed back in real time to ensure continuous supervision of risky components.

[0018] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0020] Figure 1 It is a flowchart of the steps of a method for risk supervision of a refining heating furnace provided by one embodiment of the present invention;

[0021] Figure 2 It is a system structure diagram of a risk supervision system for a refining heating furnace provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0022] The specific implementation of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the present invention, and is not used to limit the present invention.

[0023] The refining heating furnace is one of the core equipment of the refining enterprise, and its operating stability is directly related to the operating stability of the entire refining enterprise. In order to ensure the effective progress of the refining process, it is essential to continuously monitor the operating stability of the refining heating furnace.

[0024] In the existing scheme, the most common supervision scheme for refining heating furnaces is to analyze the operation data, identify the faulty equipment, inspect and troubleshoot the faulty equipment, and restart the operation after troubleshooting. This method often requires a large number of shutdowns for maintenance. During the maintenance process, the cause of the fault is reversely fuzzy located based on the abnormal data, and then the fault is troubleshooted based on the fuzzy positioning results to solve the fault problem. But in fact, in the actual operation process, although some components are abnormal, they have a certain ability to continue working, such as simple equipment aging. The existing scheme often qualitatively analyzes whether there is a fault in the equipment, and it cannot quantitatively evaluate the risk status of the equipment. For the scheme that can still work continuously, it can only be directly replaced or continuously operated based on experience. As for when to replace in order to maximize the benefits of cost control, the existing scheme is obviously unable to meet the requirements.

[0025] In view of the problem that the existing risk supervision scheme for refining and heating furnaces cannot quantitatively supervise risks, the scheme of the present invention proposes a new risk supervision method for refining and heating furnaces. The scheme of the present invention first comprehensively collects the operating data of the refining and heating furnace, and then performs data preprocessing to retain valid data and reduce the data volume. Then, data classification is performed based on the risk type to facilitate subsequent differentiated risk analysis. The risk level of the risk type is determined based on the specific data analysis results, and the supervision scheme is called based on the risk level to achieve differentiated supervision of the risk level and quantitative differentiation of risks. Then, risks are continuously supervised based on the supervision scheme, and the supervision results are fed back in real time to ensure continuous supervision of risky components.

[0026] Figure 1 This is a method flow chart of a method for risk supervision of a refining heating furnace provided by an embodiment of the present invention. Figure 1 As shown, an embodiment of the present invention provides a method for risk supervision of a refining heating furnace, the method comprising:

[0027] Step S10: collecting real-time operation data of the refining heating furnace, and preprocessing the real-time operation data to obtain preprocessed data.

[0028] Specifically, the scheme of the present invention is also based on the real-time operation information of the refining heating furnace to monitor the operation status of the refining heating furnace. In order to ensure full monitoring of the operation status of the refining heating furnace, it is necessary to comprehensively collect its operation information. The operation information includes the operation information of the equipment itself and the monitoring information of other installed monitoring equipment. The refining heating furnace does not play a role alone in the refining enterprise. Its operation status is related to the operation status of other systems. Therefore, in order to achieve full perception of the state of the refining heating furnace, it is also necessary to collect the operation data of other related systems.

[0029] Specifically, as mentioned above, in order to realize the full state perception of the refining heating furnace, it is necessary to collect its own operating data, monitoring equipment data and operating data of other related systems. If risk analysis is directly based on such a large amount of data, the amount of data processing will become very large. Then the corresponding risk judgment delay will be higher, which is contrary to the extremely fast risk supervision response speed. Therefore, before conducting a specific risk analysis, it is necessary to reduce the data volume and reduce the data processing volume to reduce the risk analysis delay.

[0030] Embodiment 1:

[0031] Furthermore, after collecting the operating data, the scheme of the present invention will pre-process the operating data, including filtering the real-time operating data based on the component type of the refining heating furnace, retaining valid data as the first data set; parsing the data in the first data set to unify the format of all data, and using the parsed data as the second data set; de-dimensionalizing the data in the second data set to obtain evaluation data, and using the evaluation data as the third data set.

[0032] In the embodiment of the present invention, the method of reducing the data volume of the present invention is to retain valid data and filter out invalid data. Because the operation data of many other related devices are collected synchronously, although there is a linkage relationship with the refining heating furnace, only part of the data is linked, and a large amount of data only exists in the internal linkage.

[0033] Specifically, for each component type, a data topology tree of the corresponding component type is extracted; based on the leaf nodes in the data topology tree, valid data is extracted one by one in the real-time operation data; all component types and all real-time operation data are traversed to obtain all valid data, and the remaining data is filtered.

[0034] Preferably, the method also includes constructing a data topology tree for each component type, including: using the operating data directly related to each component type as the parent node data; using the related data that directly affects the parent node data as the child node data of the corresponding parent node, thereby extending the nodes to form an initial topology tree; in the initial topology tree, identifying the in-degree of all nodes, and after retaining any branch, deleting other branches of the nodes with in-degree greater than 1 to obtain a second topology tree; in the second topology tree, identifying the data of all nodes to determine whether there are nodes with the same data. If so, retain the node with the highest level and delete other nodes with the same data as the node to obtain a data topology tree for each component type.

[0035] In the embodiments of the present invention, the existing data filtering scheme often directly identifies invalid data, and then filters the identified invalid data, and the retained data is defaulted to valid data. Because the amount of data collected by the scheme of the present invention is too large, there may be more invalid data than valid data. Based on this, when the scheme of the present invention performs data filtering, valid data is identified, and then the remaining data is defaulted to invalid data for filtering.

[0036] Embodiment 2:

[0037] Furthermore, when valid data is identified, the scheme of the present invention performs data identification based on a topological tree. For a certain risk type, such as a furnace nozzle risk type, the directly related operating data is the flame state information of the furnace nozzle. However, the flame state information is affected by the fuel supply and the combustion-supporting material supply. When the flame state is used as the parent node, its corresponding child nodes are the fuel supply data and the combustion-supporting material supply data. The corresponding fuel supply data is determined by information such as the fuel gas type, air pressure and concentration, so it is necessary to further extend the topological tree. Based on the topological tree rules, starting from the source, the data of the corresponding leaf nodes can be extracted one by one, and the extracted data can be used as valid data. By analogy, all valid data can be obtained and the remaining data can be filtered.

[0038] Furthermore, the in-degrees of all nodes are identified, and after retaining any branch, other branches of the nodes with in-degree greater than 1 are deleted to obtain a second topological tree.

[0039] In the embodiment of the present invention, in the topological tree, in order to ensure the validity of the data and avoid invalid branches, it is necessary to ensure that all nodes have only one parent node, so as to avoid the secondary use of the same data in subsequent analysis, resulting in inaccurate analysis results. Some data may affect multiple data, which means that the data will become the child node data of multiple data, resulting in multiple in-degrees of the node. Because the subsequent analysis is of the entire risk type, in this case, one of the branches can be retained and the other branches can be deleted. In the subsequent analysis, it is sufficient to ensure that the data only appears once in the data set.

[0040] Furthermore, the data of all nodes are identified to determine whether there are nodes with the same data. If so, the node with the highest level is retained, and other nodes with the same data as the node are deleted to obtain the data topology tree of each component type.

[0041] Similarly, when a parent node determines a child node, it may be the same as the child nodes of other nodes, but in this case there may be inconsistent levels. In this case, if the branch cannot be deleted, you can delete nodes based on the same node. While retaining the highest level node, delete subsequent nodes, which can also avoid the problem of repeatedly extracting the same data.

[0042] Furthermore, during the data collection process, the data sources are different and the corresponding data formats may be different. If all the data need to be analyzed later, it is necessary to ensure that all data formats are unified to avoid problems such as unrecognizable data or data fusion. Based on this, the solution of the present invention also needs to unify the data formats and convert all data into data of the same format.

[0043] Furthermore, during the operation of the equipment, there are various data, such as the flame state, which is information such as the area and height of the flame, and the corresponding fuel supply, which is information such as flow rate and valve opening degree. These data types, dimensions and meanings are all different. When performing numerical processing later, these data cannot be processed directly. Based on this, the scheme of the present invention will also perform dedimensionalization processing to ensure that the representation meaning of all data is consistent. Preferably, any dedimensionalization method among data standardization, normalization, proportional standard method and regularization can be adopted.

[0044] Step S20: classify the preprocessed data based on risk type to obtain data sets under each risk type.

[0045] Embodiment 3:

[0046] Specifically, based on the component type information involved in each risk type, the preprocessed data is classified based on the determined component information to obtain the preprocessed data of all component types under each risk type; based on the preset level of each risk type, the preprocessed data under each risk type is sorted, and the higher the preset level of the risk type, the higher the corresponding data is sorted; the classification result after sorting is used as the classification result to obtain the data set under each risk type.

[0047] In an embodiment of the present invention, the risk types include any of furnace tube risk type, furnace nozzle risk type, furnace structure risk type, and monitoring equipment risk type. There are multiple related components under each risk type, such as image monitoring equipment, vibration monitoring equipment, temperature monitoring equipment, etc. under the monitoring equipment risk type. The scheme of the present invention needs to differentiate risk data based on different types, so that when the risk level is determined later, parallel processing can be performed based on the differentiated data. Because before determining the specific risk level, it is impossible to guarantee which part of the data has a high risk level, and the order of judgment cannot be determined. In order to avoid delaying the processing of data with a high priority and causing an increase in delay, data differentiation is performed here, which can achieve parallel processing of data and ensure that the risk levels of all risk types can be output synchronously.

[0048] Step S30: Determine the risk level of the corresponding risk type based on the data set under each risk type one by one, and call the corresponding supervision plan in turn according to the determined risk level based on the preset risk knowledge graph.

[0049] Specifically, based on the data sets under each risk type, corresponding risk type data simulation is performed to obtain simulation results; the simulation results of each risk type are compared with a pre-built normal refining and heating furnace simulation model to obtain a deviation value; a weighted operation is performed based on the preset level of each risk type and the deviation value to obtain a priority evaluation score corresponding to each risk type; wherein the weight of the preset level of each risk type and the weight corresponding to the deviation value are both obtained based on a preset weight table; the priority evaluation scores of each risk type are compared, and the higher the score, the higher the corresponding risk level.

[0050] In an embodiment of the present invention, a corresponding level is preset for each risk type, and the higher the level, the greater the degree of impact. For example, for the monitoring equipment risk type and the furnace nozzle risk type, after the former fails, its impact is only the inability to accurately perceive the operating status in the furnace, but other equipment in the furnace does not necessarily fail, so the operating status of the corresponding refining heating furnace may not change. If the furnace nozzle fails, it will directly affect the heating of the medium in the furnace tube, which will directly affect the stability of the entire refining enterprise. Based on this, the preset level of the furnace nozzle failure type should be higher than the preset level of the monitoring equipment risk type.

[0051] The present invention performs simulation model training based on operating data. Having completed data differentiation in advance, the present invention can then perform synchronous simulation model construction for each risk type. Furthermore, since each simulation model construction is performed locally, the corresponding data processing volume will not be too large, and the model construction can be completed quickly.

[0052] Embodiment 4:

[0053] Furthermore, a weighted operation is performed on the preset level based on each risk type and the deviation value to obtain the priority evaluation score corresponding to each risk type. As mentioned above, if there is the same deviation, the score of the furnace nozzle risk type should be greater than the score of the monitoring equipment risk type. In order to reflect this difference, the scheme of the present invention presets a weight table, determines the weight to be used by the deviation value and the preset level, and then calculates the score based on the weight. The dedimensionalization process has been completed in advance, so the meaning of the deviation value here is the same, and the data operation can be directly performed.

[0054] Preferably, the method of calling a corresponding regulatory scheme for a determined risk level based on a preset risk knowledge graph includes: starting from the risk type with the highest risk level, locating the risk in the risk type, and obtaining risk component information; searching the preset risk knowledge graph based on the risk component information to obtain a regulatory scheme corresponding to the current risk component; and traversing all risk types based on risk level sorting to obtain regulatory schemes corresponding to all risk components.

[0055] In an embodiment of the present invention, the higher the risk level, the higher the priority of supervision, and limited execution processing is required. The solution of the present invention calls the supervision solution based on the risk level to achieve differentiated supervision and improve supervision performance.

[0056] Step S40: Execute the called supervision plan and output corresponding feedback information based on the execution result.

[0057] Specifically, for the located risk components, redundant collection of corresponding risk component monitoring data is performed while ensuring the same regular monitoring data collection as other components of the same type; wherein the redundant collection includes increasing the sampling frequency or extending the sampling period; the regular collection data and the redundant collection data are used together as the monitoring data of the risk components, and the real-time monitoring of the operating status of the risk components is performed based on the monitoring data.

[0058] Preferably, if the real-time monitoring result of the risk component shows that it is continuously in a risky state, monitoring data reports are continuously generated, and the corresponding data reports are fed back to the regulatory end once every predetermined period; if the real-time monitoring result of the risk component shows that it is out of the risky state, the regulatory plan for the risk component is stopped, and the corresponding out-of-risk signal is output to the regulatory end; if the real-time monitoring result of the risk component shows that its risk level has increased, alarm information is continuously output and pushed to the regulatory end.

[0059] In a possible implementation, when monitoring the flame status of the furnace nozzle, for key monitoring parts (risk components), in addition to normal image recognition, a secondary identification result is added, that is, after the abnormal flame combustion is determined, 5 consecutive frames of pictures are immediately selected to verify the flame combustion status. For areas where the burner or furnace tube abnormalities have been extinguished multiple times, the system will highlight the mark to remind on-site personnel to pay special attention. A feedback adjustment system is added to the furnace nozzle of key concern to automatically adjust the fuel gas and air supply flow. If it cannot be adjusted to the normal value, it is determined that the furnace nozzle is damaged or blocked, and a prompt to replace the furnace nozzle is given.

[0060] Embodiment 5:

[0061] Figure 2 1 is a system structure diagram of a risk supervision system for a refining and heating furnace provided by an embodiment of the present invention. Figure 2As shown, an embodiment of the present invention provides a refining heating furnace risk supervision system, the system comprising:

[0062] The collection unit is used to collect the real-time operation data of the refining heating furnace and pre-process the real-time operation data to obtain the pre-processed data.

[0063] Specifically, the scheme of the present invention is also based on the real-time operation information of the refining heating furnace to monitor the operation status of the refining heating furnace. In order to ensure full monitoring of the operation status of the refining heating furnace, it is necessary to comprehensively collect its operation information. The operation information includes the operation information of the equipment itself and the monitoring information of other installed monitoring equipment. The refining heating furnace does not play a role alone in the refining enterprise. Its operation status is related to the operation status of other systems. Therefore, in order to achieve full perception of the state of the refining heating furnace, it is also necessary to collect the operation data of other related systems.

[0064] Specifically, as mentioned above, in order to realize the full state perception of the refining heating furnace, it is necessary to collect its own operating data, monitoring equipment data and operating data of other related systems. If risk analysis is directly based on such a large amount of data, the amount of data processing will become very large. Then the corresponding risk judgment delay will be higher, which is contrary to the extremely fast risk supervision response speed. Therefore, before conducting a specific risk analysis, it is necessary to reduce the data volume and reduce the data processing volume to reduce the risk analysis delay.

[0065] Furthermore, after collecting the operating data, the scheme of the present invention will pre-process the operating data, including filtering the real-time operating data based on the component type of the refining heating furnace, retaining valid data as the first data set; parsing the data in the first data set to unify the format of all data, and using the parsed data as the second data set; de-dimensionalizing the data in the second data set to obtain evaluation data, and using the evaluation data as the third data set.

[0066] In the embodiment of the present invention, the method of reducing the data volume of the present invention is to retain valid data and filter out invalid data. Because the operation data of many other related devices are collected synchronously, although there is a linkage relationship with the refining heating furnace, only part of the data is linked, and a large amount of data only exists in the internal linkage.

[0067] Specifically, for each component type, a data topology tree of the corresponding component type is extracted; based on the leaf nodes in the data topology tree, valid data is extracted one by one in the real-time operation data; all component types and all real-time operation data are traversed to obtain all valid data, and the remaining data is filtered.

[0068] Preferably, the method also includes constructing a data topology tree for each component type, including: using the operating data directly related to each component type as the parent node data; using the related data that directly affects the parent node data as the child node data of the corresponding parent node, thereby extending the nodes to form an initial topology tree; in the initial topology tree, identifying the in-degree of all nodes, and after retaining any branch, deleting other branches of the nodes with in-degree greater than 1 to obtain a second topology tree; in the second topology tree, identifying the data of all nodes to determine whether there are nodes with the same data. If so, retain the node with the highest level and delete other nodes with the same data as the node to obtain a data topology tree for each component type.

[0069] In the embodiments of the present invention, the existing data filtering scheme often directly identifies invalid data, and then filters the identified invalid data, and the retained data is defaulted to valid data. Because the amount of data collected by the scheme of the present invention is too large, there may be more invalid data than valid data. Based on this, when the scheme of the present invention performs data filtering, valid data is identified, and then the remaining data is defaulted to invalid data for filtering.

[0070] Furthermore, when valid data is identified, the scheme of the present invention performs data identification based on a topological tree. For a certain risk type, such as a furnace nozzle risk type, the directly related operating data is the flame state information of the furnace nozzle. However, the flame state information is affected by the fuel supply and the combustion-supporting material supply. When the flame state is used as the parent node, its corresponding child nodes are the fuel supply data and the combustion-supporting material supply data. The corresponding fuel supply data is determined by information such as the fuel gas type, air pressure and concentration, so it is necessary to further extend the topological tree. Based on the topological tree rules, starting from the source, the data of the corresponding leaf nodes can be extracted one by one, and the extracted data can be used as valid data. By analogy, all valid data can be obtained and the remaining data can be filtered.

[0071] Furthermore, the in-degrees of all nodes are identified, and after retaining any branch, other branches of the nodes with in-degree greater than 1 are deleted to obtain a second topological tree.

[0072] In the embodiment of the present invention, in the topological tree, in order to ensure the validity of the data and avoid invalid branches, it is necessary to ensure that all nodes have only one parent node, so as to avoid the secondary use of the same data in subsequent analysis, resulting in inaccurate analysis results. Some data may affect multiple data, which means that the data will become the child node data of multiple data, resulting in multiple in-degrees of the node. Because the subsequent analysis is of the entire risk type, in this case, one of the branches can be retained and the other branches can be deleted. In the subsequent analysis, it is sufficient to ensure that the data only appears once in the data set.

[0073] Furthermore, the data of all nodes are identified to determine whether there are nodes with the same data. If so, the node with the highest level is retained, and other nodes with the same data as the node are deleted to obtain the data topology tree of each component type.

[0074] Similarly, when a parent node determines a child node, it may be the same as the child nodes of other nodes, but in this case there may be inconsistent levels. In this case, if the branch cannot be deleted, you can delete nodes based on the same node. While retaining the highest level node, delete subsequent nodes, which can also avoid the problem of repeatedly extracting the same data.

[0075] Furthermore, during the data collection process, the data sources are different and the corresponding data formats may be different. If all the data need to be analyzed later, it is necessary to ensure that all data formats are unified to avoid problems such as unrecognizable data or data fusion. Based on this, the solution of the present invention also needs to unify the data formats and convert all data into data of the same format.

[0076] Furthermore, during the operation of the equipment, there are various data, such as the flame state, which is information such as the area and height of the flame, and the corresponding fuel supply, which is information such as flow rate and valve opening degree. These data types, dimensions and meanings are all different. When performing numerical processing later, these data cannot be processed directly. Based on this, the scheme of the present invention will also perform dedimensionalization processing to ensure that the representation meaning of all data is consistent. Preferably, any dedimensionalization method among data standardization, normalization, proportional standard method and regularization can be adopted.

[0077] The processing unit is used to classify the pre-processed data based on risk types to obtain data sets under each risk type.

[0078] Specifically, based on the component type information involved in each risk type, the preprocessed data is classified based on the determined component information to obtain the preprocessed data of all component types under each risk type; based on the preset level of each risk type, the preprocessed data under each risk type is sorted, and the higher the preset level of the risk type, the higher the corresponding data is sorted; the classification result after sorting is used as the classification result to obtain the data set under each risk type.

[0079] In an embodiment of the present invention, the risk types include any of furnace tube risk type, furnace nozzle risk type, furnace structure risk type, and monitoring equipment risk type. There are multiple related components under each risk type, such as image monitoring equipment, vibration monitoring equipment, temperature monitoring equipment, etc. under the monitoring equipment risk type. The scheme of the present invention needs to differentiate risk data based on different types, so that when the risk level is determined later, parallel processing can be performed based on the differentiated data. Because before determining the specific risk level, it is impossible to guarantee which part of the data has a high risk level, and the order of judgment cannot be determined. In order to avoid delaying the processing of data with a high priority and causing an increase in delay, data differentiation is performed here, which can achieve parallel processing of data and ensure that the risk levels of all risk types can be output synchronously.

[0080] The scheme calling unit is used to determine the risk level of the corresponding risk type based on the data set under each risk type one by one, and based on the preset risk knowledge graph, call the corresponding supervision scheme in turn according to the determined risk level.

[0081] Specifically, based on the data sets under each risk type, corresponding risk type data simulation is performed to obtain simulation results; the simulation results of each risk type are compared with a pre-built normal refining and heating furnace simulation model to obtain a deviation value; a weighted operation is performed based on the preset level of each risk type and the deviation value to obtain a priority evaluation score corresponding to each risk type; wherein the weight of the preset level of each risk type and the weight corresponding to the deviation value are both obtained based on a preset weight table; the priority evaluation scores of each risk type are compared, and the higher the score, the higher the corresponding risk level.

[0082] In an embodiment of the present invention, a corresponding level is preset for each risk type, and the higher the level, the greater the degree of impact. For example, for the monitoring equipment risk type and the furnace nozzle risk type, after the former fails, its impact is only the inability to accurately perceive the operating status in the furnace, but other equipment in the furnace does not necessarily fail, so the operating status of the corresponding refining heating furnace may not change. If the furnace nozzle fails, it will directly affect the heating of the medium in the furnace tube, which will directly affect the stability of the entire refining enterprise. Based on this, the preset level of the furnace nozzle failure type should be higher than the preset level of the monitoring equipment risk type.

[0083] The present invention performs simulation model training based on operating data. Having completed data differentiation in advance, the present invention can then perform synchronous simulation model construction for each risk type. Furthermore, since each simulation model construction is performed locally, the corresponding data processing volume will not be too large, and the model construction can be completed quickly.

[0084] Furthermore, a weighted operation is performed on the preset level based on each risk type and the deviation value to obtain the priority evaluation score corresponding to each risk type. As mentioned above, if there is the same deviation, the score of the furnace nozzle risk type should be greater than the score of the monitoring equipment risk type. In order to reflect this difference, the scheme of the present invention presets a weight table, determines the weight to be used by the deviation value and the preset level, and then calculates the score based on the weight. The dedimensionalization process has been completed in advance, so the meaning of the deviation value here is the same, and the data operation can be directly performed.

[0085] Preferably, the method of calling a corresponding regulatory scheme for a determined risk level based on a preset risk knowledge graph includes: starting from the risk type with the highest risk level, locating the risk in the risk type, and obtaining risk component information; searching the preset risk knowledge graph based on the risk component information to obtain a regulatory scheme corresponding to the current risk component; and traversing all risk types based on risk level sorting to obtain regulatory schemes corresponding to all risk components.

[0086] In an embodiment of the present invention, the higher the risk level, the higher the priority of supervision, and limited execution processing is required. The solution of the present invention calls the supervision solution based on the risk level to achieve differentiated supervision and improve supervision performance.

[0087] The execution unit is used to execute the called supervision plan and output corresponding feedback information based on the execution result.

[0088] Specifically, for the located risk components, redundant collection of corresponding risk component monitoring data is performed while ensuring the same regular monitoring data collection as other components of the same type; wherein the redundant collection includes increasing the sampling frequency or extending the sampling period; the regular collection data and the redundant collection data are used together as the monitoring data of the risk components, and the real-time monitoring of the operating status of the risk components is performed based on the monitoring data.

[0089] Preferably, if the real-time monitoring result of the risk component shows that it is continuously in a risky state, monitoring data reports are continuously generated, and the corresponding data reports are fed back to the regulatory end once every predetermined period; if the real-time monitoring result of the risk component shows that it is out of the risky state, the regulatory plan for the risk component is stopped, and the corresponding out-of-risk signal is output to the regulatory end; if the real-time monitoring result of the risk component shows that its risk level has increased, alarm information is continuously output and pushed to the regulatory end.

[0090] An embodiment of the present invention further provides a computer-readable storage medium, on which instructions are stored, which, when executed on a computer, enable the computer to execute the above-mentioned refining heating furnace risk supervision method.

[0091] Those skilled in the art will understand that all or part of the steps in the method for implementing the above-mentioned embodiments can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including several instructions for making a single-chip microcomputer, a chip or a processor (processor) perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0092] The optional embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the technical concept of the embodiments of the present invention, the technical scheme of the embodiments of the present invention can be subjected to a variety of simple modifications, and these simple modifications all belong to the protection scope of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the embodiments of the present invention will not further describe various possible combinations.

[0093] In addition, various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the embodiments of the present invention, they should also be regarded as the contents disclosed in the embodiments of the present invention.

Claims

1. A method for risk supervision of a refining heating furnace, characterized in that: The method comprises: Collecting real-time operation data of the refining heating furnace, and preprocessing the real-time operation data to obtain preprocessed data; Classifying the preprocessed data based on risk types to obtain data sets under each risk type; Determine the risk level of the corresponding risk type based on the data set under each risk type one by one, and call the corresponding supervision plan in turn according to the determined risk level based on the preset risk knowledge graph; Execute the called supervision plan and output corresponding feedback information based on the execution results.

2. The method according to claim 1, characterized in that Preprocessing the real-time operation data includes: Based on the component type of the refining heating furnace, the real-time operation data is filtered and valid data is retained as a first data set; Parsing the data in the first data set to unify the format of all data, and using the parsed data as the second data set; De-dimensionalization is performed on the data in the second data set to obtain evaluation data, and the evaluation data is used as the third data set.

3. The method according to claim 2, characterized in that The filtering of the real-time operation data based on the component type of the refining heating furnace includes: For each component type, extract the data topology tree of the corresponding component type; Based on the leaf nodes in the data topology tree, extracting effective data one by one in the real-time operation data; Traverse all component types and all real-time operation data, obtain all valid data, and filter the remaining data.

4. The method according to claim 3, characterized in that The method further comprises: constructing a data topology tree of each component type, including: The operating data directly related to each component type is used as the parent node data; The relevant data that directly affects the parent node data is used as the child node data of the corresponding parent node, so as to extend the node and form the initial topology tree; In the initial topological tree, identify the in-degree of all nodes, and delete other branches after retaining any branch for nodes with in-degree greater than 1 to obtain the second topological tree; In the second topological tree, the data of all nodes are identified to determine whether there are nodes with the same data. If so, the node with the highest level is retained and other nodes with the same data as the node are deleted to obtain the data topological tree of each component type.

5. The method according to claim 1, characterized in that The classifying the pre-processed data based on risk type includes: Based on the component types involved in each risk type, the preprocessed data is classified to obtain the preprocessed data of all component types under each risk type; Based on the preset level of each risk type, the pre-processed data under each risk type is sorted. The higher the preset level of the risk type, the higher the corresponding data is sorted. Based on the sorted classification results, the data sets under each risk type are obtained.

6. The method according to claim 1, characterized in that The risk level of each risk type is determined based on the data sets under each risk type, including: Based on the data sets under each risk type, corresponding risk type data simulation is performed to obtain simulation results; Compare the simulation results of each risk type with the pre-built normal refining heating furnace simulation model to obtain the deviation value; A weighted calculation is performed based on the preset level of each risk type and the deviation value to obtain a priority evaluation score corresponding to each risk type; wherein the weight of the preset level of each risk type and the weight corresponding to the deviation value are obtained based on a preset weight table; Compare the priority evaluation scores of each risk type. The higher the score, the higher the corresponding risk level.

7. The method according to claim 1, characterized in that The method of calling a corresponding supervision scheme based on a preset risk knowledge graph for a determined risk level includes: Starting from the risk type with the highest risk level, locate the risk in the risk type and obtain the risk component information; Searching the preset risk knowledge graph based on the risk component information to obtain a supervision plan corresponding to the current risk component; Traverse all risk types based on risk level sorting and obtain the corresponding supervision plans for all risk components.

8. The method according to claim 7, characterized in that The types of risks include: Any of furnace tube risk type, furnace nozzle risk type, furnace structure risk type and monitoring equipment risk type.

9. The method according to claim 8, characterized in that The supervision scheme for executing the call includes: For the located risk components, redundant monitoring data collection of the corresponding risk components is performed while ensuring the same conventional monitoring data collection as other components of the same type; The redundant acquisition includes increasing the sampling frequency or extending the sampling period; The conventionally collected data and the redundantly collected data are used together as monitoring data of the risk component, and the operating status of the risk component is monitored in real time based on the monitoring data.

10. The method according to claim 9, characterized in that The outputting corresponding feedback information based on the execution result includes: If the real-time monitoring results of the risk component show that it is continuously in a risky state, a monitoring data report will be continuously generated, and the corresponding data report will be fed back to the supervisory end once every predetermined period; If the real-time monitoring result of the risk component shows that it is out of risk state, the supervision plan of the risk component is stopped, and the corresponding out-of-risk signal is output to the supervision end; If the real-time monitoring results of the risk component show that its risk level increases, the alarm information will be continuously output and pushed to the supervision end.

11. A refining heating furnace risk supervision system, characterized in that: The system comprises: A collection unit, used to collect real-time operation data of the refining heating furnace, and pre-process the real-time operation data to obtain pre-processed data; A processing unit, configured to classify the preprocessed data based on risk types to obtain data sets under each risk type; A scheme calling unit is used to determine the risk level of the corresponding risk type based on the data sets under each risk type one by one, and to call the corresponding supervision schemes in turn for the determined risk levels based on the preset risk knowledge graph; The execution unit is used to execute the called supervision plan and output corresponding feedback information based on the execution result.

12. A computer-readable storage medium having instructions stored thereon, which, when executed on a computer, enables the computer to execute the refining heating furnace risk supervision method as described in any one of claims 1 to 10.

Images