Intelligent detection method and system for cooling tower

Abstract

The application relates to the technical field of intelligent detection, and particularly discloses a cooling tower intelligent detection method and system, which comprises the following steps: creating a tower body model, synchronously determining the detection difficulty and detection utility value of each position in the tower body model; determining a calibration path according to the tower body model, obtaining the data anomaly degree of each position in the tower body model based on the calibration path; selecting a position according to the detection difficulty, the detection utility value and the data anomaly degree, and inserting a detection module; receiving the data uploaded by the detection module, constructing a multidimensional data body, extracting abnormal features in the multidimensional data body, and performing abnormal detection on the real-time received data. The application constructs a tower body model, selects a position in the tower body model, inserts a detection module, obtains the environmental data in the tower, constructs a multidimensional data body, and performs abnormal detection on the tower environment based on the multidimensional data body. The application is complementary to the existing detection process relying on equipment, and greatly optimizes the operation process of the cooling tower.

Description

Technical Field

[0001] This invention relates to the field of intelligent detection technology, specifically to an intelligent detection method and system for cooling towers. Background Technology

[0002] A cooling tower is a device that uses water as a circulating coolant to absorb heat from the system and release it into the atmosphere to lower the water temperature. Its cooling effect is achieved by the heat exchange between water and air flow to generate steam. The steam evaporates and carries away the heat, thus achieving heat dissipation through evaporation, convection, and radiation. This process dissipates waste heat generated in industrial processes or refrigeration and air conditioning systems to lower the water temperature and ensure the normal operation of the system. The device is generally cylindrical, hence the name cooling tower.

[0003] Existing cooling tower testing processes mostly rely on detection modules built into the equipment inside the tower to obtain the working status of these devices, analyze the working status, and then determine the working status of the cooling tower. This is an indirect tower-in-the-tower testing process. In fact, the tower-in-the-tower environment is also very important. Therefore, how to provide a solution for intelligent detection of the tower-in-the-tower environment is the technical problem that this invention aims to solve. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent detection method and system for cooling towers to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for intelligent detection of cooling towers, the method comprising:

[0007] The BIM model of the tower and the equipment models within the tower are queried. A tower model is created based on the BIM model and the equipment models. The detection difficulty and detection utility value of each location are determined synchronously in the tower model. The density of the locations is a preset value. The detection difficulty represents the difficulty of installing the detection module, and the detection utility value represents the benefit brought by installing the detection module.

[0008] The calibration path is determined based on the tower model, and the data anomaly degree at each location in the tower model is obtained based on the calibration path.

[0009] Based on the detection difficulty, detection utility value, and data anomaly degree, select the location and insert the detection module;

[0010] The system receives data uploaded by the detection module, constructs a multidimensional data volume, extracts abnormal features from the multidimensional data volume, and performs anomaly detection on the data received in real time.

[0011] As a further aspect of the present invention: the steps of querying the BIM model of the tower and the equipment model in the tower, creating a tower model based on the BIM model and the equipment model, and simultaneously determining the detection difficulty and detection utility value of each location in the tower model include:

[0012] Query the BIM model of the tower;

[0013] Query the equipment models and distribution information of the equipment within the tower;

[0014] Based on the distribution information, the equipment model is inserted into the BIM model to obtain the tower model;

[0015] Locate the bearing surface in the tower model, determine the uniformly distributed points in the tower model according to the preset density, remove the points contained in the equipment model, and mark the remaining points as the locations to be analyzed.

[0016] The detection difficulty and detection utility value of each location to be analyzed are determined based on the distribution information of the equipment models in the tower model.

[0017] As a further aspect of the present invention: the step of determining the detection difficulty and detection utility value of each location to be analyzed based on the distribution information of the equipment models in the tower model includes:

[0018] For any location to be analyzed, create an expanding sphere with an ever-increasing radius centered on that location until the radius reaches a preset radius threshold.

[0019] When the extended sphere is just tangent to a certain bearing surface, the radius of the extended sphere is read. When the extended sphere is not tangent to any bearing surface during the increasing process, the radius threshold is selected as the radius of the extended sphere.

[0020] Query the final expanded sphere, calculate the intersection of the final expanded sphere with each equipment model in the tower model, and count the number of equipment model types corresponding to all intersections;

[0021] The detection difficulty is determined based on the radius of the extended sphere, and the detection utility value is determined based on the number of types.

[0022] Among them, the detection difficulty is directly proportional to the radius of the expanded sphere, and the detection utility value is directly proportional to the number of species.

[0023] As a further aspect of the present invention: the step of determining the calibration path based on the tower model and obtaining the data anomaly degree at each location in the tower model based on the calibration path includes:

[0024] Read the locations to be analyzed and generate a calibration path that passes through all locations to be analyzed;

[0025] Based on the calibration path containing the location to be analyzed, a motion command containing hovering instructions is generated and sent to the drone;

[0026] It receives detection data containing location and time tags from the detector installed on the drone;

[0027] The detection data is analyzed to obtain the data anomaly degree of each location to be analyzed.

[0028] As a further aspect of the present invention: the step of analyzing the detection data to obtain the data anomaly degree of each location to be analyzed includes:

[0029] For any location to be analyzed, statistical analysis of detection data at different times is performed.

[0030] Based on the statistically obtained detection data, a data change function is fitted. A preset number of data points are randomly selected from all the detection data as test data to verify the data change function and obtain the accuracy of each test data point.

[0031] Analyze the accuracy of all test data to determine the data anomaly at the location to be analyzed;

[0032] Once the data anomaly rate for each location to be analyzed has been calculated, the data anomaly rate for any location to be analyzed is fitted using a Gaussian kernel function to obtain the final data anomaly rate.

[0033] As a further aspect of the present invention: the steps of receiving data uploaded by the detection module, constructing a multi-dimensional data volume, extracting abnormal features from the multi-dimensional data volume, and performing anomaly detection on the real-time received data include:

[0034] Receive data uploaded by the detection module, which contains location tags, time tags, and type tags;

[0035] Using location tags, time tags, and type tags as indexes, statistical data is collected to construct a multidimensional data volume; the multidimensional data volume is a multidimensional array.

[0036] Receive abnormal areas and abnormal time periods marked by staff, and locate sub-data volumes in the multi-dimensional data volume based on the abnormal areas and abnormal time periods;

[0037] The sub-data bodies are statistically analyzed to obtain a sub-data body library, which serves as anomaly features. The statistical process includes: when the similarity between two sub-data bodies reaches a preset similarity threshold, only one sub-data body is retained.

[0038] Anomaly detection is performed on the real-time received data based on abnormal features.

[0039] The present invention also provides an intelligent detection system for cooling towers, the system comprising:

[0040] The model creation and analysis module is used to query the BIM model of the tower and the equipment models within the tower, and to create a tower model based on the BIM model and equipment models. Simultaneously, the detection difficulty and detection utility value of each location are determined within the tower model; the density of these locations is a preset value; the detection difficulty represents the difficulty of installing the detection module, and the detection utility value represents the benefits brought by installing the detection module.

[0041] The data calibration and analysis module is used to determine the calibration path based on the tower model and obtain the data anomaly degree at each location in the tower model based on the calibration path.

[0042] The location selection module is used to select a location based on the detection difficulty, detection utility value, and data anomaly degree, and then insert the detection module.

[0043] The anomaly detection module is used to receive data uploaded by the detection module, construct a multidimensional data volume, extract abnormal features from the multidimensional data volume, and perform anomaly detection on the data received in real time.

[0044] As a further aspect of the present invention: the model creation and analysis module includes:

[0045] The model query unit is used to query the BIM model of the tower.

[0046] The distribution information query unit is used to query the equipment model and distribution information of the equipment in the tower.

[0047] The model insertion unit is used to insert the equipment model into the BIM model based on the distribution information to obtain the tower model.

[0048] The location determination unit is used to locate the bearing surface in the tower model, determine the uniform distribution points in the tower model according to the preset density, remove the points contained in the equipment model, and mark the remaining points as the locations to be analyzed.

[0049] The location analysis unit is used to determine the detection difficulty and detection utility value of each location to be analyzed based on the distribution information of the equipment models in the tower model.

[0050] As a further aspect of the present invention: the data calibration and analysis module includes:

[0051] The path generation unit is used to read the locations to be analyzed and generate a calibration path that passes through all the locations to be analyzed.

[0052] The command sending unit is used to generate motion commands containing hovering commands based on the calibration path containing the location to be analyzed, and send them to the UAV;

[0053] The first data receiving unit is used to receive detection data containing location and time tags obtained by the detector installed on the drone;

[0054] Anomaly calculation unit is used to analyze the detection data and obtain the data anomaly degree of each location to be analyzed.

[0055] As a further aspect of the present invention: the anomaly detection module includes:

[0056] The second data receiving unit is used to receive data uploaded by the detection module, which contains location tags, time tags, and type tags;

[0057] A data volume construction unit is used to use location labels, time labels, and type labels as indexes to collect statistical data and construct a multidimensional data volume; the multidimensional data volume is a multidimensional array.

[0058] The sub-data body positioning unit is used to receive the abnormal area and abnormal time period marked by the staff, and to locate the sub-data body in the multi-dimensional data body according to the abnormal area and abnormal time period.

[0059] The sub-data body statistics unit is used to count sub-data bodies and obtain a sub-data body library as anomaly features. The statistics process includes: when the similarity between two sub-data bodies reaches a preset similarity threshold, only one sub-data body is retained.

[0060] The detection execution unit is used to perform anomaly detection on the real-time received data based on abnormal characteristics.

[0061] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention constructs a tower model, selects a position in the tower model, inserts a detection module to obtain environmental data inside the tower, constructs a multi-dimensional data volume, and performs anomaly detection on the environment inside the tower based on the multi-dimensional data volume. This complements the existing detection process relying on equipment and greatly optimizes the operation process of the cooling tower. Attached Figure Description

[0062] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention.

[0063] Figure 1 This is a flowchart of an intelligent detection method for cooling towers.

[0064] Figure 2 This is the first sub-flowchart of the intelligent detection method for cooling towers.

[0065] Figure 3 This is the second sub-flowchart of the intelligent detection method for cooling towers.

[0066] Figure 4 This is the third sub-flowchart of the intelligent detection method for cooling towers.

[0067] Figure 5 This is a block diagram showing the composition and structure of an intelligent detection system for cooling towers. Detailed Implementation

[0068] To make the technical problems, solutions, and beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0069] Figure 1 This is a flowchart of an intelligent detection method for cooling towers. In this embodiment of the invention, an intelligent detection method for cooling towers includes:

[0070] Step S100: Query the BIM model of the tower and the equipment model in the tower, create a tower model based on the BIM model and equipment model, and simultaneously determine the detection difficulty and detection utility value of each location in the tower model; the density of the location is a preset value;

[0071] A cooling tower is a type of building that has a building model during the design phase, namely the BIM model mentioned above. At the same time, the equipment in the cooling tower is also standardized equipment, including execution equipment such as pumps, as well as connection equipment such as pipes. Based on the scale of the BIM model, the equipment models of these devices are inserted into the BIM model to obtain the tower body model.

[0072] The tower model can be analyzed using computer equipment. Specifically, locations are selected evenly in the tower model, and each location is analyzed to determine two parameters: detection difficulty and detection utility value. Detection difficulty refers to the difficulty of installing the detection module, while detection utility value represents the benefits brought by installing the detection module.

[0073] Step S200: Determine the calibration path based on the tower model, and obtain the data anomaly degree at each location in the tower model based on the calibration path;

[0074] Further analysis of the tower model allows for the determination of a path, known as the calibration path, within the tower model. This calibration path is then sent to a drone, which acquires data from different locations within the tower model. This process occurs during the testing phase, with the drone working cyclically along the calibration path. The number of drones may not be unique. Time-domain analysis of the acquired data reveals whether the data at each location is abnormal during the tower's operation, represented by the parameter of data anomaly.

[0075] Step S300: Select a location based on the detection difficulty, detection utility value, and data anomaly degree, and insert the detection module;

[0076] After the above processing, the detection difficulty, detection utility value, and data anomaly of each location can be obtained in advance. By analyzing these three parameters, different locations can be evaluated, and better locations can be selected for inserting the detection module. The detection module is a sensor with transmission function, which works in real time in actual applications to acquire data from different locations. The detection module is a higher-level concept, and it can acquire many types of data, which are specifically set by the staff. Generally, the data types include temperature.

[0077] Regarding the selection process, staff first determine a cost, then the number of detection modules. Next, a selection score is determined based on detection difficulty, detection utility value, and data anomaly. The selection score is inversely proportional to detection difficulty, directly proportional to detection utility value, and directly proportional to data anomaly. Furthermore, the weights of the three parameters—detection difficulty, detection utility value, and data anomaly—are different to determine the importance of each parameter, i.e., its impact on the selection score. Detection modules are then selected in descending order of their selection scores and inserted.

[0078] It is worth mentioning that, in order to ensure the uniformity of the positions, after selecting a position, the selection scores of all positions within a preset range centered on that position can be set to zero, so as to ensure that there is only one selected position within that range.

[0079] Step S400: Receive data uploaded by the detection module, construct a multidimensional data volume, extract abnormal features from the multidimensional data volume, and perform anomaly detection on the data received in real time;

[0080] Steps S400 and S200 are different stages. Step S400 is the practical application stage, and step S200 is the testing stage. The data obtained in step S200 is used to install the detection module. After installation, step S400 is the stage of real-time data acquisition, receiving data uploaded by the detection module, statistically analyzing the received data, and constructing a multidimensional data volume. The multidimensional data volume is essentially multidimensional data, including at least the dimensions representing time, spatial coordinates, and type. Time uses one dimension, spatial coordinates use three dimensions, and type uses one dimension, thus obtaining a five-dimensional array. In fact, there can be other attributes, such as data differences and integrals over a period of time. These are expandable objects, and the specifics are determined by the staff as needed.

[0081] After constructing the multidimensional data volume, all data in the tower is stored in the form of an array. The sub-data within the abnormal time interval and abnormal spatial interval are queried in the multidimensional data volume and statistically analyzed as abnormal features. In practical applications, the latest time is used as a boundary, and a time boundary is determined forward by combining the preset backtracking time. The multidimensional data volume is then truncated. For the truncated data volume, the abnormal features are traversed and matched. If the match is successful, it indicates that there is an anomaly. This is the anomaly detection process.

[0082] Figure 2 The first sub-flowchart of the intelligent detection method for cooling towers includes the following steps: querying the BIM model of the tower and the equipment models within the tower; creating a tower model based on the BIM model and equipment models; and simultaneously determining the detection difficulty and detection utility value of each location within the tower model.

[0083] Step S101: Query the BIM model of the tower;

[0084] Step S102: Query the equipment model and distribution information of the equipment in the tower;

[0085] Step S103: Insert the equipment model into the BIM model according to the distribution information to obtain the tower model;

[0086] Step S104: Locate the bearing surface in the tower model, determine the uniform distribution points in the tower model according to the preset density, remove the points contained in the equipment model, and mark the remaining points as the locations to be analyzed.

[0087] Step S105: Determine the detection difficulty and detection utility value of each location to be analyzed based on the distribution information of the equipment models in the tower model.

[0088] In one example of the technical solution of this invention, the process of determining the detection difficulty and detection utility value is described. The BIM model of the tower is queried, along with the equipment models and distribution information of the devices within the tower. The distribution information indicates the location of the devices. Based on the distribution information, the equipment models are inserted into the BIM model to obtain the tower model. The bearing surfaces are located within the tower model; these bearing surfaces are the surfaces where the detection modules can be installed. The rules for installation are preset by the staff, such as requiring a flat surface and a weldable material. Each surface in the tower model has its own attributes. The process of determining the bearing surfaces is very simple; some attribute conditions are preset, and all surfaces are then filtered.

[0089] Based on the above, points are evenly distributed in the tower model according to the preset density. Points included in the equipment model are removed, and the remaining points are marked as the locations to be analyzed. This process means that the locations to be analyzed only use points not included in the equipment model. The operation process of the equipment model is determined by the monitoring module built into the equipment model. This is an existing technology. Therefore, the detection process of the equipment itself will not be described in detail in this invention.

[0090] Furthermore, the step of determining the detection difficulty and detection utility value of each location to be analyzed based on the distribution information of the equipment models in the tower model includes:

[0091] For any location to be analyzed, create an expanding sphere with an ever-increasing radius centered on that location until the radius reaches a preset radius threshold.

[0092] When the extended sphere is just tangent to a certain bearing surface, the radius of the extended sphere is read. When the extended sphere is not tangent to any bearing surface during the increasing process, the radius threshold is selected as the radius of the extended sphere.

[0093] Query the final expanded sphere, calculate the intersection of the final expanded sphere with each equipment model in the tower model, and count the number of equipment model types corresponding to all intersections;

[0094] The detection difficulty is determined based on the radius of the extended sphere, and the detection utility value is determined based on the number of types.

[0095] Among them, the detection difficulty is directly proportional to the radius of the expanded sphere, and the detection utility value is directly proportional to the number of species.

[0096] In one embodiment of the technical solution of this invention, a specific process for determining the detection difficulty and detection utility value is provided. For any position to be analyzed, an expanding sphere with an ever-increasing radius is created with that position as the center until the radius reaches a preset radius threshold. In layman's terms, this process involves creating a sphere with an ever-increasing radius. When the expanding sphere is tangent to the first receiving surface, the radius at this time is read as the radius of the expanding sphere. When the expanding sphere is not tangent to any receiving surface during the increasing process, the radius threshold is selected as the radius of the expanding sphere. During this process, the expansion process continues regardless of whether it is tangent to the receiving surface.

[0097] Then, the final extended sphere is queried, the intersection of the final extended sphere and each equipment model in the tower model is calculated, and the number of equipment models corresponding to all intersections is counted. This process is used to determine how many types of equipment are around the location. Finally, the detection difficulty is determined based on the radius of the extended sphere, and the detection utility value is determined based on the number of types.

[0098] One feasible solution is:

[0099] In the formula, Due to the difficulty of detection, To expand the radius of the sphere, The preset correction coefficient means that the detection difficulty increases with the increase of the radius of the extended sphere, but the increase in detection difficulty becomes less and less significant as the radius of the extended sphere increases, because external installation equipment is required, and the difficulty lies only in the size of the installation equipment.

[0100] The utility value can be directly taken as the number of species, since the number of species is itself an integer.

[0101] Figure 3 This is the second sub-flowchart of the intelligent detection method for cooling towers. The step of determining the calibration path based on the tower model and obtaining the data anomaly degree at each location in the tower model based on the calibration path includes:

[0102] Step S201: Read the locations to be analyzed and generate a calibration path that passes through all locations to be analyzed;

[0103] Step S202: Generate motion commands containing hovering instructions based on the calibration path containing the location to be analyzed, and send them to the UAV;

[0104] Step S203: Receive detection data containing location and time tags from the detector installed on the drone;

[0105] Step S204: Analyze the detection data to obtain the data anomaly degree of each location to be analyzed.

[0106] The process of determining data anomalies is relatively simple. It occurs during the testing phase. The locations to be analyzed are read, and a calibration path passing through all locations to be analyzed is generated. Since the locations are uniformly distributed, they can be connected in sequence. For example, the planes are continuously queried from high to low. For each plane, for example, if there are N*N points, all points are connected by row or column to obtain the path within each plane. Then, the points between planes are connected to obtain the final path, which is called the calibration path.

[0107] Furthermore, the step of analyzing the detection data to obtain the data anomaly degree at each location to be analyzed includes:

[0108] For any location to be analyzed, statistical analysis of detection data at different times is performed.

[0109] Based on the statistically obtained detection data, a data change function is fitted. A preset number of data points are randomly selected from all the detection data as test data to verify the data change function and obtain the accuracy of each test data point.

[0110] Analyze the accuracy of all test data to determine the data anomaly at the location to be analyzed;

[0111] Once the data anomaly rate for each location to be analyzed has been calculated, the data anomaly rate for any location to be analyzed is fitted using a Gaussian kernel function to obtain the final data anomaly rate.

[0112] In one example of the technical solution of this invention, the calculation process of data anomaly is described. For any location to be analyzed, detection data at different times are statistically analyzed. A data change function is fitted based on the statistically analyzed detection data. A preset number of data are randomly selected from all the detection data as test data. Then, fitted data of the selected data is generated using the fitted data change function. This fitted data is compared with the test data, and the ratio of the difference to the test data is used as the error. The inverse ratio of the error is used as the accuracy. After obtaining the accuracy of each test data, the accuracy of all test data is statistically analyzed to determine the data anomaly of the location to be analyzed. After the data anomaly of each location to be analyzed is calculated, for any location to be analyzed, the data anomaly is fitted based on the Gaussian kernel function to obtain the final data anomaly.

[0113] The process of fitting data outliers based on the Gaussian kernel function involves querying the data outliers of the surrounding locations for each position, then summing them according to the weights of the Gaussian function to obtain the data outlier at that position, which is then used as the fitted data outlier.

[0114] Figure 4 The third sub-flowchart of the intelligent detection method for cooling towers includes the following steps: receiving data uploaded by the detection module, constructing a multi-dimensional data volume, extracting abnormal features from the multi-dimensional data volume, and performing anomaly detection on the real-time received data:

[0115] Step S401: Receive data uploaded by the detection module containing location tags, time tags, and type tags;

[0116] Step S402: Using location tags, time tags, and type tags as indexes, statistical data is collected to construct a multidimensional data volume; the multidimensional data volume is a multidimensional array;

[0117] Step S403: Receive the abnormal areas and abnormal time periods marked by the staff, and locate the sub-data volumes in the multi-dimensional data volume according to the abnormal areas and abnormal time periods;

[0118] Step S404: Statistically analyze the sub-data bodies to obtain a sub-data body library, which serves as anomaly features; wherein, the statistical process includes: when the similarity between two sub-data bodies reaches a preset similarity threshold, only one sub-data body is retained;

[0119] Step S405: Perform anomaly detection on the real-time received data based on anomaly features.

[0120] As a preferred embodiment of the technical solution of the present invention, the anomaly detection process is described. Data containing location tags, time tags, and type tags uploaded by the detection module is received in real-time (with a preset data upload frequency). The location tags, time tags, and type tags are used as indexes to collect statistical data and construct a multidimensional data volume. The multidimensional data volume is a multidimensional array, which is an array with pre-defined probabilities representing multiple dimensions. Then, abnormal regions and abnormal time periods are received, as defined by staff. Abnormal regions are used to determine spatial ranges, and abnormal time periods are used to determine temporal ranges. Sub-data volumes are located within the multidimensional data volume based on the abnormal regions and abnormal time periods. These sub-data volumes are statistically analyzed to obtain a sub-data volume library, which serves as anomaly features. During the statistical analysis, if the similarity between two sub-data volumes is sufficiently high, only one sub-data volume is retained.

[0121] Finally, anomaly detection is performed on the real-time received data based on abnormal characteristics. The detection process is as follows:

[0122] Using the latest moment as one time boundary, and combining it with a preset backtracking time to determine another time boundary, the multidimensional data volume is truncated. For the truncated data volume, anomaly features are used to traverse and match it. If the match is successful, it means that there is an anomaly. If the match is unsuccessful, it means that there is no anomaly. As time goes by, the two time boundaries are constantly changing, and correspondingly, the anomaly detection process is also real-time.

[0123] Figure 5 This is a block diagram of the composition of an intelligent cooling tower detection system. In this embodiment of the invention, an intelligent cooling tower detection system 10 includes:

[0124] The model creation and analysis module 11 is used to query the BIM model of the tower and the equipment model in the tower, create a tower model based on the BIM model and the equipment model, and simultaneously determine the detection difficulty and detection utility value of each location in the tower model; the density of the location is a preset value.

[0125] The data calibration and analysis module 12 is used to determine the calibration path based on the tower model and obtain the data anomaly degree at each location in the tower model based on the calibration path.

[0126] The location selection module 13 is used to select a location based on the detection difficulty, detection utility value and data anomaly degree, and insert the detection module therein.

[0127] The anomaly detection module 14 is used to receive data uploaded by the detection module, construct a multi-dimensional data volume, extract abnormal features from the multi-dimensional data volume, and perform anomaly detection on the data received in real time.

[0128] Furthermore, the model creation and analysis module 11 includes:

[0129] The model query unit is used to query the BIM model of the tower.

[0130] The distribution information query unit is used to query the equipment model and distribution information of the equipment in the tower.

[0131] The model insertion unit is used to insert the equipment model into the BIM model based on the distribution information to obtain the tower model.

[0132] The location determination unit is used to locate the bearing surface in the tower model, determine the uniform distribution points in the tower model according to the preset density, remove the points contained in the equipment model, and mark the remaining points as the locations to be analyzed.

[0133] The location analysis unit is used to determine the detection difficulty and detection utility value of each location to be analyzed based on the distribution information of the equipment models in the tower model.

[0134] Specifically, the data calibration and analysis module 12 includes:

[0135] The path generation unit is used to read the locations to be analyzed and generate a calibration path that passes through all the locations to be analyzed.

[0136] The command sending unit is used to generate motion commands containing hovering commands based on the calibration path containing the location to be analyzed, and send them to the UAV;

[0137] The first data receiving unit is used to receive detection data containing location and time tags obtained by the detector installed on the drone;

[0138] Anomaly calculation unit is used to analyze the detection data and obtain the data anomaly degree of each location to be analyzed.

[0139] Furthermore, the anomaly detection module 14 includes:

[0140] The second data receiving unit is used to receive data uploaded by the detection module, which contains location tags, time tags, and type tags;

[0141] A data volume construction unit is used to use location labels, time labels, and type labels as indexes to collect statistical data and construct a multidimensional data volume; the multidimensional data volume is a multidimensional array.

[0142] The sub-data body positioning unit is used to receive the abnormal area and abnormal time period marked by the staff, and to locate the sub-data body in the multi-dimensional data body according to the abnormal area and abnormal time period.

[0143] The sub-data body statistics unit is used to count sub-data bodies and obtain a sub-data body library as anomaly features. The statistics process includes: when the similarity between two sub-data bodies reaches a preset similarity threshold, only one sub-data body is retained.

[0144] The detection execution unit is used to perform anomaly detection on the real-time received data based on abnormal characteristics.

[0145] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent detection of cooling towers, characterized in that, The method includes: The BIM model of the tower and the equipment models within the tower are queried. A tower model is created based on the BIM model and the equipment models. The detection difficulty and detection utility value of each location are determined synchronously in the tower model. The density of the locations is a preset value. The detection difficulty represents the difficulty of installing the detection module, and the detection utility value represents the benefit brought by installing the detection module. The calibration path is determined based on the tower model, and the data anomaly degree at each location in the tower model is obtained based on the calibration path. Based on the detection difficulty, detection utility value, and data anomaly degree, select the location and insert the detection module; The system receives data uploaded by the detection module, constructs a multidimensional data volume, extracts abnormal features from the multidimensional data volume, and performs anomaly detection on the real-time received data. The steps of querying the BIM model of the tower and the equipment model within the tower, creating a tower model based on the BIM model and equipment model, and simultaneously determining the detection difficulty and detection utility value of each location in the tower model include: Query the BIM model of the tower; Query the equipment models and distribution information of the equipment within the tower; Based on the distribution information, the equipment model is inserted into the BIM model to obtain the tower model; Locate the bearing surface in the tower model, determine the uniformly distributed points in the tower model according to the preset density, remove the points contained in the equipment model, and mark the remaining points as the locations to be analyzed. The detection difficulty and detection utility value of each location to be analyzed are determined based on the distribution information of the equipment models in the tower model; The steps for determining the detection difficulty and detection utility value of each location to be analyzed based on the distribution information of the equipment models in the tower model include: For any location to be analyzed, create an expanding sphere with an ever-increasing radius centered on that location until the radius reaches a preset radius threshold. When the extended sphere is just tangent to a certain bearing surface, the radius of the extended sphere is read. When the extended sphere is not tangent to any bearing surface during the increasing process, the radius threshold is selected as the radius of the extended sphere. Query the final expanded sphere, calculate the intersection of the final expanded sphere with each equipment model in the tower model, and count the number of equipment model types corresponding to all intersections; The detection difficulty is determined based on the radius of the extended sphere, and the detection utility value is determined based on the number of types. Among them, the detection difficulty is directly proportional to the radius of the expanded sphere, and the detection utility value is directly proportional to the number of species.

2. The intelligent detection method for cooling towers according to claim 1, characterized in that, The steps of determining the calibration path based on the tower model and obtaining the data anomaly degree at each location in the tower model based on the calibration path include: Read the locations to be analyzed and generate a calibration path that passes through all locations to be analyzed; Based on the calibration path containing the location to be analyzed, a motion command containing hovering instructions is generated and sent to the drone; It receives detection data containing location and time tags from the detector installed on the drone; The detection data is analyzed to obtain the data anomaly degree of each location to be analyzed.

3. The intelligent detection method for cooling towers according to claim 2, characterized in that, The step of analyzing the detection data to obtain the data anomaly degree at each location to be analyzed includes: For any location to be analyzed, statistical analysis of detection data at different times is performed. Based on the statistically obtained detection data, a data change function is fitted. A preset number of data points are randomly selected from all the detection data as test data to verify the data change function and obtain the accuracy of each test data point. Analyze the accuracy of all test data to determine the data anomaly at the location to be analyzed; Once the data anomaly rate for each location to be analyzed has been calculated, the data anomaly rate for any location to be analyzed is fitted using a Gaussian kernel function to obtain the final data anomaly rate.

4. The intelligent detection method for cooling towers according to claim 1, characterized in that, The steps of receiving and detecting data uploaded by the detection module, constructing a multi-dimensional data volume, extracting abnormal features from the multi-dimensional data volume, and performing anomaly detection on the real-time received data include: Receive data uploaded by the detection module, which contains location tags, time tags, and type tags; Using location tags, time tags, and type tags as indexes, statistical data is collected to construct a multidimensional data volume; the multidimensional data volume is a multidimensional array. Receive abnormal areas and abnormal time periods marked by staff, and locate sub-data volumes in the multi-dimensional data volume based on the abnormal areas and abnormal time periods; The sub-data bodies are statistically analyzed to obtain a sub-data body library, which serves as anomaly features. The statistical process includes: when the similarity between two sub-data bodies reaches a preset similarity threshold, only one sub-data body is retained. Anomaly detection is performed on the real-time received data based on abnormal features.

5. An intelligent detection system for cooling towers, characterized in that, The system includes: The model creation and analysis module is used to query the BIM model of the tower and the equipment models within the tower, and to create a tower model based on the BIM model and equipment models. Simultaneously, the detection difficulty and detection utility value of each location are determined within the tower model; the density of these locations is a preset value; the detection difficulty represents the difficulty of installing the detection module, and the detection utility value represents the benefits brought by installing the detection module. The data calibration and analysis module is used to determine the calibration path based on the tower model and obtain the data anomaly degree at each location in the tower model based on the calibration path. The location selection module is used to select a location based on the detection difficulty, detection utility value, and data anomaly degree, and then insert the detection module. The anomaly detection module is used to receive data uploaded by the detection module, construct a multidimensional data volume, extract abnormal features from the multidimensional data volume, and perform anomaly detection on the data received in real time. The model creation and analysis module includes: The model query unit is used to query the BIM model of the tower. The distribution information query unit is used to query the equipment model and distribution information of the equipment in the tower. The model insertion unit is used to insert the equipment model into the BIM model based on the distribution information to obtain the tower model. The location determination unit is used to locate the bearing surface in the tower model, determine the uniform distribution points in the tower model according to the preset density, remove the points contained in the equipment model, and mark the remaining points as the locations to be analyzed. The location analysis unit is used to determine the detection difficulty and detection utility value of each location to be analyzed based on the distribution information of the equipment models in the tower model; The steps for determining the detection difficulty and detection utility value of each location to be analyzed based on the distribution information of the equipment models in the tower model include: For any location to be analyzed, create an expanding sphere with an ever-increasing radius centered on that location until the radius reaches a preset radius threshold. When the extended sphere is just tangent to a certain bearing surface, the radius of the extended sphere is read. When the extended sphere is not tangent to any bearing surface during the increasing process, the radius threshold is selected as the radius of the extended sphere. Query the final expanded sphere, calculate the intersection of the final expanded sphere with each equipment model in the tower model, and count the number of equipment model types corresponding to all intersections; The detection difficulty is determined based on the radius of the extended sphere, and the detection utility value is determined based on the number of types. Among them, the detection difficulty is directly proportional to the radius of the expanded sphere, and the detection utility value is directly proportional to the number of species.

6. The intelligent detection system for cooling towers according to claim 5, characterized in that, The data calibration and analysis module includes: The path generation unit is used to read the locations to be analyzed and generate a calibration path that passes through all the locations to be analyzed. The command sending unit is used to generate motion commands containing hovering commands based on the calibration path containing the location to be analyzed, and send them to the UAV; The first data receiving unit is used to receive detection data containing location and time tags obtained by the detector installed on the drone; Anomaly calculation unit is used to analyze the detection data and obtain the data anomaly degree of each location to be analyzed.

7. The intelligent detection system for cooling towers according to claim 5, characterized in that, The anomaly detection module includes: The second data receiving unit is used to receive data uploaded by the detection module, which contains location tags, time tags, and type tags; A data volume construction unit is used to use location labels, time labels, and type labels as indexes to collect statistical data and construct a multidimensional data volume; the multidimensional data volume is a multidimensional array. The sub-data body positioning unit is used to receive the abnormal area and abnormal time period marked by the staff, and to locate the sub-data body in the multi-dimensional data body according to the abnormal area and abnormal time period. The sub-data body statistics unit is used to count sub-data bodies and obtain a sub-data body library as anomaly features. The statistics process includes: when the similarity between two sub-data bodies reaches a preset similarity threshold, only one sub-data body is retained. The detection execution unit is used to perform anomaly detection on the real-time received data based on abnormal characteristics.

Images