Method and device for analyzing key equipment change based on real-time sensing of dynamic working conditions

Abstract

The application discloses a key equipment change analysis method and device based on dynamic working condition real-time sensing, and mainly relates to the technical field of data processing. The method comprises the following steps: acquiring a power transmission chain of a target coal mine, combining a historical key equipment abnormal log set of the target coal mine, and constructing a multi-source potential backtracking clue library; determining a key equipment set and a key equipment working condition real-time sensing feature set; performing clue matching in the multi-source potential backtracking clue library, and determining a matched multi-source potential backtracking clue set; performing asymmetric sampling on the key equipment set and the matched multi-source potential backtracking clue set, and performing interactive trend fusion analysis on the sampling results to obtain a key equipment change analysis result set. The application has the beneficial effects that the technical problem of the existing key equipment change analysis in a coal mine being lagging and the analysis result having low fitting degree with the actual situation is solved, the real-time performance of the key equipment change analysis is improved, and the technical effect of improving the analysis reliability is achieved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a method and apparatus for analyzing changes in key equipment based on real-time perception of dynamic operating conditions. Background Technology

[0002] With the expansion of coal mine production scale and the increase in automation, the operational stability of key underground equipment (such as coal mining machines, hydraulic supports, and scraper conveyors) directly affects operational safety and production efficiency. To improve the operation and maintenance capabilities of coal mine equipment, current coal mine information systems are gradually introducing SCADA systems, DCS systems, and various types of sensors to monitor parameters such as power, current, temperature, wind speed, and gas concentration. Some systems also support the recording of historical alarms or anomaly logs, facilitating post-event review. However, existing technologies mainly remain at the level of static or isolated analysis of the status of key equipment, resulting in a relatively lagging analysis of changes in key equipment and an inability to reflect the actual situation of the equipment. Summary of the Invention

[0003] This application provides a method and apparatus for analyzing changes in key equipment based on real-time perception of dynamic operating conditions, which is used to address the technical problems in the existing technology of analyzing changes in key equipment in coal mines, which suffers from lag and low consistency between the analysis results and the actual situation.

[0004] In view of the above problems, this application provides a method and apparatus for analyzing changes in key equipment based on real-time perception of dynamic operating conditions.

[0005] The first aspect of this application provides a method for analyzing changes in key equipment based on real-time perception of dynamic operating conditions. The method includes: acquiring the power transmission chain of a target coal mine; combining this with a set of historical key equipment anomaly logs from the target coal mine to construct a multi-source potential backtracking clue database; acquiring the collaborative operation chain of the target coal mine; traversing the collaborative operation chain to perform real-time perception of equipment operating conditions, and determining a set of key equipment and a set of key equipment operating condition real-time perception features; performing clue matching in the multi-source potential backtracking clue database based on the set of key equipment operating condition real-time perception features to determine a matching set of multi-source potential backtracking clues; performing asymmetric sampling on the set of key equipment and the matching set of multi-source potential backtracking clues, and performing interactive trend fusion analysis on the sampling results to obtain a set of key equipment change analysis results.

[0006] Preferably, asymmetric sampling is performed on the set of key equipment and the set of matching multi-source potential backtracking clues, and interactive trend fusion analysis is performed on the sampling results to obtain a set of key equipment change analysis results, including: short-frequency sampling of the set of key equipment and the set of matching multi-source potential backtracking clues at a first sampling frequency to obtain a set of short-frequency sensing feature sequences of key equipment and a set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues; long-frequency sampling of the set of key equipment and the set of matching multi-source potential backtracking clues at a second sampling frequency to obtain a set of short-frequency sensing feature sequences of key equipment and a set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues; for The key equipment short-frequency sensing feature sequence set and the key equipment long-frequency sensing feature sequence set are subjected to interactive trend fusion analysis to obtain a key equipment interactive trend vector set; the key equipment short-frequency sensing feature sequence set and the key equipment long-frequency sensing feature sequence set of matching multi-source potential backtracking clues are subjected to interactive trend fusion analysis to obtain a matching multi-source potential backtracking clue interactive trend vector set; explicit trend factor analysis is performed on the key equipment interactive trend vector set, and implicit trend factor analysis is performed on the matching multi-source potential backtracking clue interactive trend vector set, and the analysis results are weighted and calculated to obtain a key equipment change analysis result set.

[0007] Preferably, the interaction trend fusion analysis of the key equipment short-frequency sensing feature sequence set and the key equipment long-frequency sensing feature sequence set to obtain a key equipment interaction trend vector set includes: performing feature change trend analysis on the key equipment short-frequency sensing feature sequence set and the key equipment long-frequency sensing feature sequence set respectively to obtain a key equipment short-frequency sensing feature trend vector set and a key equipment long-frequency sensing feature trend vector set; calculating the similarity of vector elements between the key equipment short-frequency sensing feature trend vector set and the corresponding key equipment long-frequency sensing feature trend vector set to construct an interaction trend fusion matrix set; and convolving the interaction trend fusion matrix set with the key equipment short-frequency sensing feature trend vector set respectively to obtain the matching multi-source potential backtracking clue interaction trend vector set.

[0008] Preferably, the historical anomaly monitoring duration is extracted by traversing the set of key equipment, and the minimum historical anomaly monitoring duration in the extraction results is used as the first sampling frequency, and the maximum historical anomaly monitoring duration in the extraction results is used as the second sampling frequency.

[0009] Preferably, the process involves acquiring the collaborative operation chain of the target coal mine, traversing the collaborative operation chain to perform real-time equipment condition sensing, and determining a set of key equipment and a set of key equipment condition real-time sensing features. This includes: traversing the collaborative operation chain to collect equipment condition features in real time, obtaining a set of equipment condition real-time sensing features; pre-constructing an anomaly feature analyzer, using the anomaly feature analyzer to identify anomalies in the set of equipment condition real-time sensing features, obtaining a set of key equipment condition real-time sensing features; and mapping and extracting equipment in the collaborative operation chain based on the set of key equipment condition real-time sensing features, obtaining the set of key equipment.

[0010] Preferably, the power transmission chain of the target coal mine is obtained, and a multi-source potential backtracking clue database is constructed by combining the historical critical equipment anomaly log set of the target coal mine. This includes: extracting critical equipment groups and anomaly features that appear simultaneously in the historical critical equipment anomaly log set to obtain a set of historical critical equipment groups and a set of historical anomaly features; based on the similarity between different historical anomaly features in the set of historical anomaly features, the historical critical equipment group set is aggregated to obtain Q aggregated historical critical equipment group sets and Q aggregated historical anomaly features; combining the Q aggregated historical anomaly features and the power transmission chain, the Q aggregated historical critical equipment group sets are expanded to determine Q aggregated historical critical equipment expanded sets; and the Q aggregated historical anomaly features and the Q aggregated historical critical equipment expanded sets are mapped and associated to construct the multi-source potential backtracking clue database.

[0011] Preferably, by combining the Q aggregated historical anomaly features and the power transmission chain, the set of Q aggregated historical key equipment groups is expanded to determine the expanded set of Q aggregated historical key equipment, including: searching the power transmission chain based on the feature types of the Q aggregated historical anomaly features to determine the set of Q power transmission associated equipment; and performing a union operation on the set of Q power transmission associated equipment and the set of Q aggregated historical key equipment groups to obtain the expanded set of Q aggregated historical key equipment.

[0012] A second aspect of this application provides a key equipment change analysis device based on real-time dynamic operating condition perception, the device comprising:

[0013] The system includes a multi-source potential backtracking clue database construction module, which acquires the power transmission chain of the target coal mine and combines it with the historical critical equipment anomaly log set of the target coal mine to construct a multi-source potential backtracking clue database; a real-time sensing feature set determination module, which acquires the collaborative operation chain of the target coal mine, traverses the collaborative operation chain to perform real-time equipment condition sensing, and determines the set of critical equipment and the set of real-time sensing features of critical equipment conditions; a backtracking clue determination module, which performs clue matching in the multi-source potential backtracking clue database based on the set of real-time sensing features of critical equipment conditions to determine the matching set of multi-source potential backtracking clues; and a critical equipment change analysis result set acquisition module, which performs asymmetric sampling on the set of critical equipment and the matching set of multi-source potential backtracking clues, and performs interactive trend fusion analysis on the sampling results to obtain the set of critical equipment change analysis results.

[0014] A third aspect of this application provides a computer device, the computer device comprising:

[0015] Memory, used to store executable instructions;

[0016] When the processor runs the executable instructions stored in the memory, it implements the key equipment change analysis method based on real-time perception of dynamic operating conditions described in the first aspect above.

[0017] A fourth aspect of this application provides a computer-readable storage medium storing executable instructions, characterized in that, when the executable instructions are executed by a processor, they implement the key equipment change analysis method based on real-time perception of dynamic operating conditions described in the first aspect above.

[0018] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0019] This application constructs a multi-source potential backtracking clue database by acquiring the power transmission chain of the target coal mine and combining it with the historical critical equipment anomaly logs of the target coal mine. Then, it acquires the collaborative operation chain of the target coal mine, traverses the collaborative operation chain to perform real-time equipment condition sensing, determines the set of critical equipment and the set of real-time sensing features of critical equipment conditions, and then performs clue matching in the multi-source potential backtracking clue database based on the set of real-time sensing features of critical equipment conditions to determine the matching multi-source potential backtracking clue set. Finally, it performs asymmetric sampling on the set of critical equipment and the matching multi-source potential backtracking clue set, and performs interactive trend fusion analysis on the sampling results to obtain the set of critical equipment change analysis results. This achieves the technical effect of improving the reliability and timeliness of the analysis results. Attached Figure Description

[0020] Appendix Figure 1 This is a schematic diagram of the process for analyzing changes in key equipment based on real-time perception of dynamic operating conditions, provided in an embodiment of the present invention.

[0021] Appendix Figure 2 This is a schematic diagram of the key equipment change analysis device based on real-time perception of dynamic working conditions provided in an embodiment of the present invention.

[0022] The labels shown in the attached diagram:

[0023] The module 11 is for constructing a multi-source potential backtracking clue library, the module 12 is for determining the real-time sensing feature set, the module 13 is for determining backtracking clues, and the module 14 is for obtaining the set of key equipment change analysis results. Detailed Implementation

[0024] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims. It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion; for example, a process, method, apparatus, product, or server that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices.

[0025] Example 1, as shown in the appendix Figure 1 As shown, this application provides a method for analyzing changes in key equipment based on real-time perception of dynamic operating conditions. The method includes:

[0026] Step S1: Obtain the power transmission chain of the target coal mine, and combine it with the historical critical equipment anomaly log set of the target coal mine to construct a multi-source potential backtracking clue library;

[0027] In one embodiment of this application, the power transmission chain refers to the path structure in a coal mine system where electrical energy (or other forms of energy) is transmitted from the main power supply node (such as a substation) through multiple relay devices such as frequency converters and switchgear, and finally to various power-consuming terminal devices (such as coal mining machines, scraper conveyors, etc.). It reflects the energy supply and demand relationship between devices, as well as the impact of power fluctuations at a certain node on the equipment. The historical critical equipment anomaly log set refers to the collection of historical anomaly event records related to critical equipment accumulated over long-term operation through SCADA, DCS systems, or manual recording methods. These logs typically include equipment number, anomaly time, fault description, associated equipment, environmental context, etc. The multi-source potential backtracking clue library refers to an indexed data structure that integrates power structure relationships, historical anomaly patterns, and equipment collaborative information. It is used to support the retrospective judgment and causal inference of future anomaly trends, preserving the possible paths and associated characteristics of equipment anomalies.

[0028] First, historical anomaly log data involving key equipment in the target coal mine is retrieved, and frequently co-occurring equipment groups and characteristic values ​​(such as high current, high temperature, and low wind speed) are categorized and aggregated. By analyzing the similarities and common causes among these anomalies, an aggregated set of historical key equipment groups and an aggregated set of historical anomaly features are constructed. Then, based on the power transmission chain, the power supply path of these equipment groups is extended to include upstream and downstream related equipment in the power chain, forming an expanded equipment set. Finally, the expanded equipment set is mapped and associated with its corresponding anomaly features to construct a multi-source potential backtracking clue library that supports rapid indexing, matching, and backtracking.

[0029] Furthermore, by acquiring the power transmission chain of the target coal mine and combining it with the historical critical equipment anomaly log set of the target coal mine, a multi-source potential backtracking clue database is constructed. Step S1 in this embodiment of the application also includes:

[0030] Extract the critical equipment groups and abnormal features that appear simultaneously in the historical critical equipment anomaly log set to obtain the historical critical equipment group set and the historical anomaly feature set;

[0031] Based on the similarity between different historical anomaly features in the set of historical anomaly features, the set of historical critical equipment groups is aggregated into similar groups to obtain Q aggregated sets of historical critical equipment groups and Q aggregated historical anomaly features, where Q is a positive integer;

[0032] Combining the Q aggregation historical anomaly features and power transmission chains, the set of Q aggregation historical key equipment groups is expanded to determine the expanded set of Q aggregation historical key equipment groups;

[0033] The Q aggregated historical anomaly features and the Q aggregated historical key equipment expansion sets are mapped and associated to construct the multi-source potential backtracking clue library.

[0034] Furthermore, combining the Q aggregation history anomaly features and power transmission chains, the set of Q aggregation history key equipment groups is expanded to determine the expanded set of Q aggregation history key equipment. In this embodiment, step S1 further includes:

[0035] Based on the feature types of the Q aggregated historical anomaly features, the power transmission chain is retrieved to determine a set of Q power transmission associated devices;

[0036] The union of the Q sets of power transmission associated devices and the Q sets of aggregated historical key devices is obtained to obtain the expanded set of the Q aggregated historical key devices.

[0037] In one possible embodiment, the key device groups and corresponding anomaly features that appear simultaneously in each historical key device anomaly log are traversed within the historical key device anomaly log set to obtain the historical key device group set and the historical anomaly feature set. The historical key device group set reflects the co-occurrence of devices when an anomaly occurs, providing objects for subsequent backtracking analysis of potential anomalies. The historical anomaly feature set reflects the actual situation of the devices when the anomaly occurred, providing a basis for subsequent retrieval and matching of corresponding clues.

[0038] Preferably, a first historical anomaly feature is randomly extracted from the historical anomaly feature set without replacement. Then, the similarity between the first historical anomaly feature and each other in the historical anomaly feature set is calculated using the cosine similarity formula. When the calculation result is greater than or equal to a similarity threshold preset by those skilled in the art, the corresponding historical anomaly feature is added to the first initial aggregated historical anomaly feature set without replacement. Then, a second historical anomaly feature is randomly extracted from the historical anomaly feature set again without replacement. Then, the similarity between the second historical anomaly feature and each remaining historical anomaly feature in the historical anomaly feature set is calculated using the cosine similarity formula. When the calculation result is greater than or equal to a similarity threshold preset by those skilled in the art, the corresponding historical anomaly feature is added to the second initial aggregated historical anomaly feature set without replacement, and so on. This achieves the goal of aggregating historical anomaly feature sets according to the degree of similarity between them, obtaining the Q initial aggregated historical anomaly feature sets.

[0039] In one possible embodiment, Q aggregated historical anomaly features are obtained by calculating the mean of the Q initial aggregated historical anomaly feature sets. These Q aggregated historical anomaly features are the features that best reflect the characteristics of the Q initial aggregated historical anomaly feature sets. Furthermore, based on the one-to-one correspondence between historical anomaly features and historical critical equipment groups, the historical critical equipment group set is aggregated according to the Q initial aggregated historical anomaly feature sets to obtain Q aggregated historical critical equipment group sets.

[0040] Furthermore, a physical network of power transmission chains is introduced. For the aggregated features among the Q aggregated historical anomalies, the relevant upstream and downstream devices are found in the power chain by identifying their feature types (such as power type, temperature rise type, voltage type, etc.). These related devices are then aggregated to form an expanded set of Q aggregated historical key devices associated with the key devices.

[0041] Each set of aggregated historical anomaly features is mapped one-to-one with its corresponding aggregated historical key equipment expansion set, and written into a structured database or graph engine to build a multi-source potential backtracking clue library.

[0042] By preserving the empirical paths in the anomaly logs and introducing potential influencing factor devices through power chain extension, the technical effect of providing a clue index structure with both knowledge support and physical correlation is achieved for subsequent anomaly trend analysis and causal reasoning based on sensing data.

[0043] Step S2: Obtain the collaborative operation chain of the target coal mine, traverse the collaborative operation chain to perform real-time equipment condition perception, and determine the set of key equipment and the set of key equipment condition real-time perception features.

[0044] Furthermore, the collaborative operation chain of the target coal mine is obtained, and the equipment operating conditions are traversed in real time to determine the set of key equipment and the set of key equipment operating conditions real-time sensing features. In this embodiment, step S2 further includes:

[0045] The collaborative operation chain is traversed to collect equipment operating condition characteristics in real time, and a set of real-time perceived equipment operating condition characteristics is obtained.

[0046] A pre-built anomaly feature analyzer is used to identify anomalies in the real-time sensing feature set of equipment operating conditions to obtain the real-time sensing feature set of key equipment operating conditions.

[0047] Based on the real-time sensing feature set of the key equipment operating conditions, the equipment in the collaborative operation chain is mapped and extracted to obtain the set of key equipment.

[0048] In one possible embodiment, the collaborative operation chain is a functional flow chain composed of various equipment according to operational logic in the coal mine production process, such as coal mining machine → transfer machine → scraper conveyor → main conveyor belt → loading system, etc. This chain reflects the operational dependencies between equipment, that is, the working status of upstream equipment often affects the operational stability of downstream equipment. The real-time equipment condition sensing feature set is a multi-dimensional data set that reflects the current operating status of the equipment, collected by various sensors, and may include current, voltage, vibration, temperature, torque, speed, load, etc. The key equipment set is the set of equipment in the collaborative operation chain that has experienced anomalies.

[0049] Preferably, the anomaly feature analyzer refers to a pre-built functional module for identifying potential abnormal behavior patterns from real-time equipment operating condition sensing features. Multiple sample sets of real-time equipment operating condition sensing features are acquired, and multiple sets of key sample real-time equipment operating condition sensing features exhibiting anomalies are extracted from these sets as training data. The framework built on a feedforward neural network is then trained under supervised supervision using these multiple sample sets and the multiple sets of key sample real-time equipment operating condition sensing features until training converges, resulting in the trained anomaly feature analyzer. Furthermore, the anomaly feature analyzer is used to identify anomalies in the real-time equipment operating condition sensing feature sets to obtain the key equipment operating condition real-time sensing feature sets.

[0050] Based on the identified anomaly characteristics, the system reverse-engineers the device nodes exhibiting these characteristics within the collaborative workflow chain, identifying these devices as critical equipment, thereby obtaining the set of critical equipment. This achieves the technical effect of identifying equipment requiring focused attention and its corresponding anomalies.

[0051] Step S3: Based on the real-time sensing feature set of the key equipment operating conditions, perform clue matching in the multi-source potential backtracking clue database to determine the matching multi-source potential backtracking clue set;

[0052] In one embodiment, the set of real-time perceived features of key equipment operating conditions is compared with the aggregated historical anomaly features corresponding to each multi-source potential backtracking clue in the multi-source potential backtracking clue library. Optionally, the cosine similarity calculation formula can be used for similarity calculation, and the multi-source potential backtracking clue corresponding to the maximum value of the calculation result is taken as the clue corresponding to each real-time perceived feature of key equipment operating conditions, thus obtaining the set of matching multi-source potential backtracking clues. This rapidly maps real-time perceived abnormal equipment behavior to a historical clue structure with prior meaning, establishing a connection between the current behavior and known evolutionary patterns, thereby improving the technical effect of subsequent traceability reliability.

[0053] Step S4: Perform asymmetric sampling on the set of key equipment and the set of matching multi-source potential backtracking clues, and perform interactive trend fusion analysis on the sampling results to obtain a set of key equipment change analysis results.

[0054] In one possible embodiment, the set of key equipment and the set of potential backtracking clues from multiple sources are taken as the analysis objects, and asymmetric sampling is performed on them respectively. Short-frequency sampling is used to detect rapidly fluctuating response characteristics (such as load jumps, current oscillations, etc.), while long-frequency sampling is used to identify slowly changing trends (such as long-term temperature rises, voltage drops, gas accumulation, etc.). This asynchronous sampling mechanism can be adaptively adjusted by setting two sets of sampling window periods (e.g., 10 seconds and 10 minutes) or based on historical monitoring durations.

[0055] Furthermore, trend analysis is performed on the feature sequences obtained from short-frequency and long-frequency sampling to obtain trend vectors. The similarity between the obtained trend vectors is then quantitatively evaluated to form a trend fusion matrix. This matrix is ​​then convolved with the short-frequency trend vectors to generate a fused trend vector, i.e., a set of interactive trend vectors. Next, explicit trend factor analysis (for extracting observable and clearly defined trend features) is performed on the interactive trend vectors from the key equipment side, while implicit trend factor analysis (for identifying empirical or potential path patterns) is performed on the interactive trend vectors from the clue side. The results of the two types of factors are then fused using a weighted mechanism to obtain the final set of key equipment change analysis results, which may include equipment number, trend score, trend direction, and associated clue ID. This achieves the technical effect of structured prediction of the future state evolution of key equipment and improves the timeliness of change analysis response.

[0056] Furthermore, asymmetric sampling is performed on the set of key equipment and the set of matching multi-source potential backtracking clues, and interactive trend fusion analysis is performed on the sampling results to obtain a set of key equipment change analysis results. Step S4 of this embodiment also includes:

[0057] Short-frequency sampling is performed on the set of key equipment and the set of matching multi-source potential backtracking clues according to the first sampling frequency to obtain the set of short-frequency sensing feature sequences of key equipment and the set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues;

[0058] Long-frequency sampling is performed on the set of key equipment and the set of matching multi-source potential backtracking clues according to the second sampling frequency to obtain the set of short-frequency sensing feature sequences of key equipment and the set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues;

[0059] An interactive trend fusion analysis is performed on the short-frequency sensing feature sequence set and the long-frequency sensing feature sequence set of the key equipment to obtain the key equipment interactive trend vector set.

[0060] An interactive trend fusion analysis is performed on the set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues and the set of long-frequency sensing feature sequences of matching multi-source potential backtracking clues to obtain a set of interactive trend vectors for matching multi-source potential backtracking clues.

[0061] Explicit trend factor analysis is performed on the set of interaction trend vectors of key equipment, implicit trend factor analysis is performed on the set of interaction trend vectors of matching multi-source potential backtracking clues, and the analysis results are weighted to obtain the set of key equipment change analysis results.

[0062] Furthermore, an interaction trend fusion analysis is performed on the short-frequency sensing feature sequence set and the long-frequency sensing feature sequence set of the key equipment to obtain a set of key equipment interaction trend vectors. Step S4 in this embodiment further includes:

[0063] The feature change trend analysis was performed on the short-frequency sensing feature sequence set and the long-frequency sensing feature sequence set of key equipment respectively to obtain the short-frequency sensing feature trend vector set and the long-frequency sensing feature trend vector set of key equipment.

[0064] Calculate the similarity of vector elements between the set of short-frequency sensing feature trend vectors of the key equipment and the corresponding set of long-frequency sensing feature trend vectors of the key equipment, and construct an interactive trend fusion matrix set.

[0065] The interaction trend fusion matrix set and the key equipment short-frequency perception feature trend vector set are convolved respectively to obtain the matching multi-source potential backtracking clue interaction trend vector set.

[0066] Furthermore, the historical anomaly monitoring duration is extracted by traversing the set of key equipment, and the minimum historical anomaly monitoring duration in the extraction results is used as the first sampling frequency, and the maximum historical anomaly monitoring duration in the extraction results is used as the second sampling frequency.

[0067] In one possible embodiment, to improve the rationality and dynamic adaptability of the sampling strategy, before asymmetric sampling, the historical anomaly monitoring duration of each device in the set of key devices is extracted and analyzed. This process, based on the anomaly log database or SCADA system historical curve data extracted in step S1, finds the duration of the interval between two adjacent anomalies for each key device in historical anomaly events. Then, the minimum historical anomaly monitoring duration in the set of historical anomaly monitoring durations is used as the first sampling frequency, i.e., for high-frequency short-window sampling, focusing on short-time response characteristics, and the maximum historical anomaly monitoring duration is used as the second sampling frequency, i.e., for low-frequency long-window sampling, focusing on the extraction of gradual trends.

[0068] This adaptive sampling frequency strategy can configure the granularity based on the experience window of actual device abnormal response. It avoids missing details due to overly coarse sampling granularity, while also avoiding high computational load and redundant features affecting the accuracy of trend modeling due to overly fine sampling granularity.

[0069] In one possible embodiment, explicit trend factors refer to trend factors that can be directly observed and quantified in key equipment data, such as indicators calculated from abrupt change slopes and rates of change (e.g., ∆current / ∆t). Implicit trend factors refer to the magnitude of abnormal trends analyzed from associated equipment data. Data is collected at short time intervals from the key equipment set and the matching multi-source potential backtracking clue set at a first sampling frequency, forming two short-frequency sensing feature sequence sets: the key equipment short-frequency sensing feature sequence set and the matching multi-source potential backtracking clue short-frequency sensing feature sequence set. The key equipment short-frequency sensing feature sequence set represents the rapid state changes of the current equipment, while the matching multi-source potential backtracking clue short-frequency sensing feature sequence set reflects the rapid evolution of associated equipment.

[0070] Preferably, the two objects are sampled again at a second sampling frequency to form two sets of long-frequency sensing feature sequences: a set of short-frequency sensing feature sequences for key equipment and a set of short-frequency sensing feature sequences for matching multi-source potential backtracking clues. These sets are used to record slowly changing trend features. This dual-frequency synchronous extraction mechanism ensures that the data has both real-time responsiveness and retains sufficient contextual trend information.

[0071] A trend vector recognizer is used to analyze the feature change trends of the short-frequency sensing feature sequence set and the long-frequency sensing feature sequence set of the key equipment, respectively, to obtain the trend vector sets of the short-frequency sensing features and the long-frequency sensing features of the key equipment. Preferably, multiple sample key equipment sensing feature sequences and multiple sample key equipment sensing feature trend vectors are used as training data to supervise the training of the framework built on the feedforward neural network until convergence is achieved, thus obtaining the trained trend vector recognizer.

[0072] Furthermore, using the cosine similarity calculation formula, the vector element similarity between the set of short-frequency sensing feature trend vectors of key equipment and the corresponding set of long-frequency sensing feature trend vectors of key equipment is calculated. The calculated vector element similarity set is then normalized using the min-max formula, and the processed result is added to an empty matrix to obtain the interaction trend fusion matrix set. Next, a graph convolutional network is used to perform a one-to-one mapping convolution between the interaction trend fusion matrix set and the set of short-frequency sensing feature trend vectors of key equipment, thereby obtaining the set of interaction trend vectors matching multi-source potential backtracking clues.

[0073] Based on the same principle as obtaining the set of interaction trend vectors for the key equipment, an interaction trend fusion analysis is performed on the set of short-frequency sensing feature sequences of the matching multi-source potential backtracking clues and the set of long-frequency sensing feature sequences of the matching multi-source potential backtracking clues to obtain a set of interaction trend vectors for the matching multi-source potential backtracking clues.

[0074] In one possible embodiment, a trend factor analyzer is obtained, and the trend factor analyzer is used to analyze the set of interaction trend vectors of key equipment and the set of interaction trend vectors of matched multi-source potential backtracking clues, respectively, to obtain a set of explicit trend factors and a set of implicit trend factors. According to weights preset by those skilled in the art, the set of explicit trend factors and the set of implicit trend factors are mapped and weighted to obtain a set of key equipment change analysis results.

[0075] Preferably, multiple sample trend vectors and corresponding sample trend factors are obtained as analyzer training data. The analyzer training data is divided into a training set and a validation set according to a pre-defined ratio by those skilled in the art. The training set is used to perform supervised training on the framework constructed based on a convolutional neural network. Multiple sample trend vectors from the validation set are input into the framework to obtain multiple output trend factors. The ratio between the multiple output trend factors and the multiple sample trend factors in the validation set is statistically analyzed. If the statistical results meet the requirements, the analysis is passed, and the trained trend factor analyzer is obtained.

[0076] Through explicit analysis and associated implicit analysis, the technical effect of improving the reliability and timeliness of critical equipment change analysis was achieved.

[0077] In summary, the embodiments of this application have at least the following technical effects:

[0078] This application constructs a multi-source potential backtracking clue database by acquiring the power transmission chain of the target coal mine and combining it with the historical critical equipment anomaly logs of the target coal mine. Then, it acquires the collaborative operation chain of the target coal mine, traverses the collaborative operation chain to perform real-time equipment condition sensing, determines the set of critical equipment and the set of real-time sensing features of critical equipment conditions, and then performs clue matching in the multi-source potential backtracking clue database based on the set of real-time sensing features of critical equipment conditions to determine the matching multi-source potential backtracking clue set. Finally, it performs asymmetric sampling on the set of critical equipment and the matching multi-source potential backtracking clue set, and performs interactive trend fusion analysis on the sampling results to obtain the set of critical equipment change analysis results. This achieves the technical effect of improving the reliability and timeliness of the analysis results.

[0079] Example 2 is based on the same inventive concept as the key equipment change analysis method based on real-time dynamic operating condition perception in the foregoing examples, as shown in the appendix. Figure 2 As shown, this application provides a key equipment change analysis device based on real-time dynamic working condition perception. The device and method embodiments in this application are based on the same inventive concept. The device includes:

[0080] The multi-source potential backtracking clue database construction module 11 is used to obtain the power transmission chain of the target coal mine and, in combination with the historical key equipment anomaly log set of the target coal mine, construct a multi-source potential backtracking clue database.

[0081] The real-time sensing feature set determination module 12 is used to obtain the collaborative operation chain of the target coal mine, traverse the collaborative operation chain to perform real-time sensing of equipment operating conditions, and determine the key equipment set and the key equipment operating condition real-time sensing feature set.

[0082] The backtracking clue determination module 13 is used to match clues in the multi-source potential backtracking clue database based on the real-time sensing feature set of the key equipment operating conditions, and determine the matching multi-source potential backtracking clue set.

[0083] The module 14 for obtaining the set of key equipment change analysis results is used to perform asymmetric sampling on the set of key equipment and the set of matching multi-source potential backtracking clues, and to perform interactive trend fusion analysis on the sampling results to obtain the set of key equipment change analysis results.

[0084] Furthermore, the key equipment change analysis result set acquisition module 14 is used to perform the following steps:

[0085] Short-frequency sampling is performed on the set of key equipment and the set of matching multi-source potential backtracking clues according to the first sampling frequency to obtain the set of short-frequency sensing feature sequences of key equipment and the set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues;

[0086] Long-frequency sampling is performed on the set of key equipment and the set of matching multi-source potential backtracking clues according to the second sampling frequency to obtain the set of short-frequency sensing feature sequences of key equipment and the set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues;

[0087] An interactive trend fusion analysis is performed on the short-frequency sensing feature sequence set and the long-frequency sensing feature sequence set of the key equipment to obtain the key equipment interactive trend vector set.

[0088] An interactive trend fusion analysis is performed on the set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues and the set of long-frequency sensing feature sequences of matching multi-source potential backtracking clues to obtain a set of interactive trend vectors for matching multi-source potential backtracking clues.

[0089] Explicit trend factor analysis is performed on the set of interaction trend vectors of key equipment, implicit trend factor analysis is performed on the set of interaction trend vectors of matching multi-source potential backtracking clues, and the analysis results are weighted to obtain the set of key equipment change analysis results.

[0090] Furthermore, the key equipment change analysis result set acquisition module 14 is used to perform the following steps:

[0091] The feature change trend analysis was performed on the short-frequency sensing feature sequence set and the long-frequency sensing feature sequence set of key equipment respectively to obtain the short-frequency sensing feature trend vector set and the long-frequency sensing feature trend vector set of key equipment.

[0092] Calculate the similarity of vector elements between the set of short-frequency sensing feature trend vectors of the key equipment and the corresponding set of long-frequency sensing feature trend vectors of the key equipment, and construct an interactive trend fusion matrix set.

[0093] The interaction trend fusion matrix set and the key equipment short-frequency perception feature trend vector set are convolved respectively to obtain the matching multi-source potential backtracking clue interaction trend vector set.

[0094] Furthermore, the historical anomaly monitoring duration is extracted by traversing the set of key equipment, and the minimum historical anomaly monitoring duration in the extraction results is used as the first sampling frequency, and the maximum historical anomaly monitoring duration in the extraction results is used as the second sampling frequency.

[0095] Furthermore, the real-time perceived feature set determination module 12 is used to perform the following steps:

[0096] The collaborative operation chain is traversed to collect equipment operating condition characteristics in real time, and a set of real-time perceived equipment operating condition characteristics is obtained.

[0097] A pre-built anomaly feature analyzer is used to identify anomalies in the real-time sensing feature set of equipment operating conditions to obtain the real-time sensing feature set of key equipment operating conditions.

[0098] Based on the real-time sensing feature set of the key equipment operating conditions, the equipment in the collaborative operation chain is mapped and extracted to obtain the set of key equipment.

[0099] Furthermore, the multi-source potential backtracking clue library construction module 11 is used to perform the following steps:

[0100] Extract the critical equipment groups and abnormal features that appear simultaneously in the historical critical equipment anomaly log set to obtain the historical critical equipment group set and the historical anomaly feature set;

[0101] Based on the similarity between different historical anomaly features in the set of historical anomaly features, the set of historical critical equipment groups is aggregated into similar groups to obtain Q aggregated sets of historical critical equipment groups and Q aggregated historical anomaly features, where Q is a positive integer;

[0102] Combining the Q aggregation historical anomaly features and power transmission chains, the set of Q aggregation historical key equipment groups is expanded to determine the expanded set of Q aggregation historical key equipment groups;

[0103] The Q aggregated historical anomaly features and the Q aggregated historical key equipment expansion sets are mapped and associated to construct the multi-source potential backtracking clue library.

[0104] Furthermore, the multi-source potential backtracking clue library construction module 11 is used to perform the following steps:

[0105] Based on the feature types of the Q aggregated historical anomaly features, the power transmission chain is retrieved to determine a set of Q power transmission associated devices;

[0106] The union of the Q sets of power transmission associated devices and the Q sets of aggregated historical key devices is obtained to obtain the expanded set of the Q aggregated historical key devices.

[0107] Example 3: Based on the same inventive concept as the key equipment change analysis method based on real-time dynamic working condition perception in the foregoing examples, this application provides a computer device, the computer device comprising:

[0108] Memory, used to store executable instructions;

[0109] When the processor runs the executable instructions stored in the memory, it implements the above-mentioned method for analyzing changes in key equipment based on real-time perception of dynamic operating conditions.

[0110] In some embodiments, the computer device may also optionally include an input interface and an output interface. The processor, memory, and input / output interfaces can be connected via a bus or signal lines. Various peripheral devices can be connected to the input / output interfaces via buses, signal lines, or circuit boards. The processor, having signal processing capabilities, can be a microprocessor or any conventional processor; it can be a general-purpose processor, a central processing unit (CPU), a network processor (NP), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), etc. The memory can be volatile memory or non-volatile memory, or may include both.

[0111] Example 4: Based on the same inventive concept as the key equipment change analysis method based on real-time dynamic operating condition perception in the foregoing embodiments, this application provides a computer-readable storage medium storing executable instructions. The executable instructions, when executed by a processor, implement the aforementioned key equipment change analysis method based on real-time dynamic operating condition perception. The computer-readable storage medium includes: permanent and non-permanent, removable and non-removable media, which are tangible devices capable of retaining and storing instructions for use by an instruction execution device. The computer-readable storage medium includes: electronic storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, and any suitable combination thereof.

[0112] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0113] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0114] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for analyzing changes in key equipment based on real-time dynamic operating condition perception, characterized in that: The method includes: Obtain the power transmission chain of the target coal mine, and combine it with the historical critical equipment anomaly log set of the target coal mine to construct a multi-source potential backtracking clue library. The power transmission chain reflects the energy supply and demand relationship between equipment and the equipment status affected by power fluctuations between nodes. The collaborative operation chain of the target coal mine is obtained, and the equipment operating conditions are traversed in real time to determine the set of key equipment and the set of key equipment operating conditions real time perception features. The collaborative operation chain reflects the operation dependency relationship between equipment and the operation sequence between nodes. Based on the real-time sensing feature set of the key equipment operating conditions, clue matching is performed in the multi-source potential backtracking clue database to determine the matching multi-source potential backtracking clue set. Non-uniform sampling is performed on the set of key equipment and the set of matching multi-source potential backtracking clues, and interactive trend fusion analysis is performed on the sampling results to obtain a set of key equipment change analysis results. Specifically, non-uniform sampling is performed on the set of key equipment and the set of matched multi-source potential backtracking clues, and interactive trend fusion analysis is conducted on the sampling results to obtain a set of key equipment change analysis results, including: Short-frequency sampling is performed on the set of key equipment and the set of matching multi-source potential backtracking clues according to the first sampling frequency to obtain the set of short-frequency sensing feature sequences of key equipment and the set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues; Long-frequency sampling is performed on the set of key equipment and the set of matching multi-source potential backtracking clues according to the second sampling frequency to obtain the set of long-frequency sensing feature sequences of key equipment and the set of long-frequency sensing feature sequences of matching multi-source potential backtracking clues. An interactive trend fusion analysis is performed on the short-frequency sensing feature sequence set and the long-frequency sensing feature sequence set of the key equipment to obtain the key equipment interactive trend vector set. An interactive trend fusion analysis is performed on the set of short-frequency sensing feature sequences of matching multi-source potential backtracking clues and the set of long-frequency sensing feature sequences of matching multi-source potential backtracking clues to obtain a set of interactive trend vectors for matching multi-source potential backtracking clues. Explicit trend factor analysis is performed on the set of interaction trend vectors of key equipment, implicit trend factor analysis is performed on the set of interaction trend vectors of matching multi-source potential backtracking clues, and the analysis results are weighted to obtain the set of key equipment change analysis results.

2. The method for analyzing changes in key equipment based on real-time dynamic operating condition perception as described in claim 1, characterized in that, An interaction trend fusion analysis is performed on the short-frequency sensing feature sequence set and the long-frequency sensing feature sequence set of the key equipment to obtain a set of interaction trend vectors for the key equipment, including: The feature change trend analysis was performed on the short-frequency sensing feature sequence set and the long-frequency sensing feature sequence set of key equipment respectively to obtain the short-frequency sensing feature trend vector set and the long-frequency sensing feature trend vector set of key equipment. Calculate the similarity of vector elements between the set of short-frequency sensing feature trend vectors of the key equipment and the corresponding set of long-frequency sensing feature trend vectors of the key equipment, and construct an interactive trend fusion matrix set. The interaction trend fusion matrix set and the key equipment short-frequency perception feature trend vector set are convolved respectively to obtain the matching multi-source potential backtracking clue interaction trend vector set.

3. The method for analyzing changes in key equipment based on real-time dynamic operating condition perception as described in claim 2, characterized in that, The historical anomaly monitoring duration is extracted by traversing the set of key equipment, and the minimum historical anomaly monitoring duration in the extraction results is used as the first sampling frequency, and the maximum historical anomaly monitoring duration in the extraction results is used as the second sampling frequency.

4. The method for analyzing changes in key equipment based on real-time dynamic operating condition perception as described in claim 1, characterized in that, Obtain the collaborative operation chain of the target coal mine, traverse the collaborative operation chain to perform real-time equipment condition sensing, and determine the set of key equipment and the set of key equipment condition real-time sensing features, including: The collaborative operation chain is traversed to collect equipment operating condition characteristics in real time, and a set of real-time perceived equipment operating condition characteristics is obtained. A pre-built anomaly feature analyzer is used to identify anomalies in the real-time sensing feature set of equipment operating conditions to obtain the real-time sensing feature set of key equipment operating conditions. Based on the real-time sensing feature set of the key equipment operating conditions, the equipment in the collaborative operation chain is mapped and extracted to obtain the set of key equipment.

5. The method for analyzing changes in key equipment based on real-time dynamic operating condition perception as described in claim 1, characterized in that, Obtain the power transmission chain of the target coal mine, and combine it with the historical critical equipment anomaly logs of the target coal mine to construct a multi-source potential backtracking clue database, including: Extract the critical equipment groups and abnormal features that appear simultaneously in the historical critical equipment anomaly log set to obtain the historical critical equipment group set and the historical anomaly feature set. Based on the similarity between different historical anomaly features in the set of historical anomaly features, the set of historical critical equipment groups is aggregated into similar groups to obtain Q aggregated sets of historical critical equipment groups and Q aggregated historical anomaly features, where Q is a positive integer; Combining the Q aggregation historical anomaly features and power transmission chains, the set of Q aggregation historical key equipment groups is expanded to determine the expanded set of Q aggregation historical key equipment groups; The Q aggregated historical anomaly features and the Q aggregated historical key equipment expansion sets are mapped and associated to construct the multi-source potential backtracking clue library.

6. The method for analyzing changes in key equipment based on real-time dynamic operating condition perception as described in claim 5, characterized in that, Based on the aforementioned Q aggregation historical anomaly characteristics and power transmission chains, the set of Q aggregation historical critical equipment groups is expanded to determine the expanded set of Q aggregation historical critical equipment groups, including: Based on the feature types of the Q aggregated historical anomalies, the power transmission chain is retrieved to determine a set of Q power transmission associated devices. The union of the Q sets of power transmission associated devices and the Q sets of aggregated historical key devices is obtained to obtain the expanded set of the Q aggregated historical key devices.

7. A key equipment change analysis device based on real-time dynamic operating condition perception, characterized in that, The apparatus is used to implement the key equipment change analysis method based on real-time dynamic working condition perception as described in any one of claims 1-6, and the apparatus comprises: The multi-source potential backtracking clue database construction module is used to obtain the power transmission chain of the target coal mine and combine it with the historical key equipment anomaly log set of the target coal mine to construct the multi-source potential backtracking clue database. The power transmission chain reflects the energy supply and demand relationship between equipment and the equipment status affected by power fluctuations between nodes. The real-time sensing feature set determination module is used to obtain the collaborative operation chain of the target coal mine, traverse the collaborative operation chain to perform real-time sensing of equipment operating conditions, and determine the key equipment set and the key equipment operating condition real-time sensing feature set. The collaborative operation chain reflects the operation dependency relationship between equipment and the operation sequence between nodes. The backtracking clue determination module is used to match clues in the multi-source potential backtracking clue database based on the real-time sensing feature set of the key equipment's operating conditions, and determine the matching multi-source potential backtracking clue set. The module for obtaining the set of key equipment change analysis results is used to perform non-uniform sampling on the set of key equipment and the set of matching multi-source potential backtracking clues, and to perform interactive trend fusion analysis on the sampling results to obtain the set of key equipment change analysis results.

8. A computer device, characterized in that, The computer device includes: Memory, used to store executable instructions; The processor, when running the executable instructions stored in the memory, implements the key equipment change analysis method based on real-time dynamic operating condition perception as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing executable instructions, characterized in that, When the executable instructions are executed by the processor, they implement the key equipment change analysis method based on real-time perception of dynamic operating conditions as described in any one of claims 1 to 6.

Images