Cable crane working status identification method based on massive monitoring data

Abstract

The present invention belongs to the field of pattern recognition and relates to a method for identifying the operating status of a cable crane, comprising the following steps: 1) acquiring cable crane operation monitoring data; 2) preprocessing the cable crane operation monitoring data; 3) calculating the rate of change of the cable crane's three-dimensional features based on the preprocessed cable crane operation monitoring data; 4) determining the trend state of the cable crane based on the rate of change of the cable crane's three-dimensional features; 5) using the trend state of the cable crane to perform single-cycle division of the measured cable crane data; 6) matching identification parameters of the single cycles and identifying four operating modes of the cable crane based on the identification parameters; and 7) matching the three-dimensional features based on the identification results to identify the operating status of the cable crane. The present invention provides a cable crane operation mode identification method based on massive cable crane operation monitoring data, which can effectively identify the cable crane's operation mode and operating status. The identification results provide a reliable basis for cable crane efficiency analysis.

Description

Technical Field

[0001] The present invention belongs to the field of pattern recognition, and relates to a method for identifying the working state of a cable crane, and in particular to a method for identifying the working state of a cable crane based on massive monitoring data. Background Art

[0002] Cable cranes are essential construction machinery used in concrete pouring operations for hydraulic engineering projects. Due to their large load capacity and high transport efficiency, they are crucial for concrete pouring strength, project duration, and quality. Consequently, in actual construction, cable cranes are highly valued and place high demands on their operation.

[0003] With the advancement of the theory of intelligent dam construction in practice, a new direction has been pointed out for achieving refined management and control of the construction process, personnel and equipment, and project quality and safety. Information perception is the foundation of intelligent dam construction, and quickly, timely, and accurately obtaining the working status and operating efficiency of the cable crane is an important prerequisite for the safe operation of the cable crane and the optimization and improvement of its efficiency. In order to grasp the actual operation of the cable crane, a monitoring system was jointly developed with the manufacturer and operator to achieve all-weather synchronous monitoring of the entire cable crane operation process, with the characteristics of real-time, accuracy, and high refresh rate. However, the monitoring data obtained in the current cable crane positioning monitoring system is relatively primitive, and the relevant data and information required for the calculation of the cable crane operation efficiency needs to be analyzed and extracted. To this end, pattern recognition is required to identify and analyze these monitoring data and mine information to accurately calculate relevant information such as cable crane efficiency. Summary of the Invention

[0004] In order to solve the above-mentioned technical problems existing in the background technology, the present invention provides a cable crane operation mode recognition method based on massive cable crane operation monitoring data, which can effectively identify the cable crane operation mode and working status, and its recognition results provide a reliable basis for the efficiency analysis of the cable crane.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A cable crane working state identification method based on massive monitoring data, characterized in that the cable crane working state identification method based on massive monitoring data comprises the following steps:

[0007] 1) Obtain cable crane operation monitoring data;

[0008] 2) Preprocessing the cable crane operation monitoring data obtained in step 1);

[0009] 3) Calculate the rate of change of the three-dimensional characteristics of the cable crane based on the pre-processed cable crane operation monitoring data;

[0010] 4) determining the trend state of the cable crane according to the rate of change of the three-dimensional characteristics of the cable crane obtained in step 3);

[0011] 5) using the trend state of the cable crane determined in step 4) to perform single-cycle division on the measured data of the cable crane;

[0012] 6) Match the identification parameters of the single cycle in step 5), and identify the four working modes of the cable crane based on the identification parameters.

[0013] Preferably, the specific implementation method of step 1) adopted by the present invention is: using SQL statements to extract required data from the database of the cable crane information monitoring system.

[0014] Preferably, the preprocessing in step 2) of the present invention is to delete time-duplicate data, delete erroneous data, verify and process coordinate data, convert the format of time data, and perform noise reduction on the cable crane operation monitoring data.

[0015] Preferably, the specific implementation of step 3) adopted by the present invention is:

[0016] 3.1) Obtaining pre-processed cable crane operation monitoring data, wherein the pre-processed cable crane operation monitoring data includes the cable crane station number, cable crane coordinates at the time of data collection, state duration, collection time, and cable crane hoisting weight;

[0017] 3.2) extracting the three-dimensional features of the cable crane from the pre-processed cable crane operation monitoring data; the three-dimensional features of the cable crane include the cable crane hoisting weight and spatial coordinates;

[0018] 3.3) Calculate the change rates of the three-dimensional characteristics of the cable crane in step 3.2) respectively; the change rates of the three-dimensional characteristics of the cable crane include the change rate of the lifting weight, the change rate of the horizontal displacement, and the change rate of the cable crane lifting depth.

[0019] Preferably, the method for calculating the rate of change of the hanging weight in step 3.3) of the present invention is:

[0020]

[0021] The calculation method of the horizontal displacement change rate is:

[0022]

[0023] The calculation method of the cable crane hoisting depth change rate is:

[0024]

[0025] in:

[0026] is the first-order difference value of the lifting weight, that is, the rate of change of the lifting weight;

[0027] W nis the weight of the nth row;

[0028] W n+1 It is the weight of the n+1th row;

[0029] ΔW is the difference in lifting weight between the front and rear rows;

[0030] is the first-order difference of the horizontal displacement, that is, the rate of change of the horizontal displacement;

[0031] X n is the horizontal displacement of the nth row;

[0032] X n+1 is the horizontal displacement of the n+1th row;

[0033] ΔX is the horizontal displacement difference between the two rows;

[0034] is the first-order difference of vertical displacement, that is, the rate of change of vertical displacement;

[0035] Z n is the vertical displacement of the nth row;

[0036] Z n+1 is the vertical displacement of row n+1;

[0037] ΔZ is the vertical displacement difference between the two rows;

[0038] t n is the acquisition time of the nth row;

[0039] t n+1 is the acquisition time of row n+1;

[0040] Δt is the time difference between the two lines.

[0041] Preferably, the specific implementation of step 4) adopted by the present invention is:

[0042] 4.1) Setting the change threshold;

[0043] 4.2) Determine whether the rate of change of the three-dimensional characteristics of the cable crane is within the change threshold set in step 4.1). If so, the cable crane is considered to be in a stable state; if not, proceed to step 4.3);

[0044] 4.3) Determine whether the cable crane is in an ascending or descending state based on the positive or negative rate of change of the cable crane's three-dimensional characteristics; a positive rate of change indicates that the cable crane is in an ascending state; a negative rate of change indicates that the cable crane is in a descending state.

[0045] Preferably, step 4.3) adopted in the present invention further includes:

[0046] 4.4) Aggregate single time points with the same state according to the trend of the three-dimensional feature change. After aggregation, calculate the mean of the feature values ​​within the state interval. After the aggregation state, the initial time of the state is retained; this is used to determine the trend of the three-dimensional feature change between two adjacent states.

[0047] Preferably, the specific implementation of step 5) adopted by the present invention is:

[0048] 5.1) Perform boundary identification based on the trend state identification results at the above single time point;

[0049] 5.2) Numbering the results of step 5.1) in ascending order;

[0050] 5.3) When the trend state of the cable car appears continuously, identify its starting point and end point;

[0051] 5.4) Taking the starting time of the trend state of the cable crane as the starting point, complete the single-cycle division of the cable crane measured data.

[0052] Preferably, the identification parameters in step 6) adopted in the present invention include the number of heavy-load hoisting times, hoisting weight characteristics, and time characteristics;

[0053] The four working modes of the cable crane are: concrete lifting mode, miscellaneous object lifting mode, maintenance mode and shutdown mode;

[0054] The identification parameters of the concrete lifting mode are: the number of heavy-load lifting is 2 times; the lifting weight is not higher than 28 tons on the forward trip and not lower than 6.5 tons on the return trip;

[0055] The identification parameters of the lifting mode of sporadic objects are: the number of heavy-load lifting is 1; the lifting weight characteristic is 1 to 30 tons;

[0056] The identification parameters of the maintenance mode are: the number of heavy-load lifting is 0 times; the lifting weight characteristic is 0 to 1 t; the time characteristic is 1 hour;

[0057] The identification parameters of the shutdown mode are: the number of heavy-load lifting is 0 times; the lifting weight characteristic is 0 to 1 t; and the time characteristic is ≥5 minutes.

[0058] Preferably, the cable crane working status identification method based on massive monitoring data adopted by the present invention further includes, after step 6):

[0059] 7) Based on the recognition results of step 6), the three-dimensional features are matched to identify the working status of the cable crane within a single cycle; the cable crane working status includes waiting for loading, lifting and ascending, transportation preparation, horizontal transportation, descending into the warehouse, warehouse surface alignment, unloading, warehouse out of the warehouse, ascending out of the warehouse, empty tank return, platform alignment, descending back to the original position, maintenance status, and shutdown status.

[0060] The advantages of the present invention are:

[0061] The present invention discloses a method for identifying the operating status of a cable crane based on massive monitoring data, comprising: 1) obtaining cable crane operation monitoring data from a database; 2) preprocessing the cable crane operation monitoring data obtained in step 1); 3) performing first-order difference calculations based on the characteristic values ​​of the preprocessed data in step 2), and calculating the characteristic change rate; 4) setting a threshold to determine the trend state; 5) dividing single cycles based on the trend state in step 4); 6) identifying four operating modes of the cable crane based on matching single cycle identification parameters in step 5); and 7) matching three-dimensional features based on the identification results in step 6) to identify the operating status of the cable crane. The present invention is a method for identifying, analyzing, and mining relatively raw massive monitoring data from current cable crane positioning and monitoring systems to accurately extract relevant information, such as calculating cable crane efficiency. The present invention utilizes pattern recognition technology to extract a large number of attribute characteristic values, and uses their regularity to perform feature classification. This method extracts relevant information for various efficiency analyses from the massive monitoring data, thereby identifying the operating mode and operating status of the cable crane. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a flow chart of the cable crane working status identification method based on massive monitoring data provided by the present invention;

[0063] Figure 2 The cable crane operation mode used in the present invention is divided into four modes;

[0064] Figure 3 This is a structural diagram of the concrete lifting mode adopted by the present invention;

[0065] Figure 4 This is a structural diagram of the lifting of scattered objects adopted by the present invention;

[0066] Figure 5 This is a flow chart for identifying the working status of a cable crane used in the present invention;

[0067] Figure 6 The present invention uses the method of extracting the original cable crane operation data graph from the database;

[0068] Figure 7 This is the state aggregation and feature value recognition result diagram adopted by the present invention;

[0069] Figure 8 This is a diagram of the single-cycle recognition result of the operating mode adopted by the present invention;

[0070] Figure 9 This is the working status recognition result diagram adopted by the present invention. DETAILED DESCRIPTION

[0071] See also Figure 1The present invention provides a method for identifying the working status of a cable crane based on massive monitoring data, the method comprising the following steps:

[0072] Step 1: Obtain cable crane operation monitoring data from the database, specifically:

[0073] The cable crane information monitoring system database contains approximately 16 million records of monitoring data for each cable crane from April 10, 2017, to December 15, 2020. SQL statements are used to extract relevant data as needed.

[0074] Step 2: Preprocess the acquired cable crane operation monitoring data, specifically:

[0075] Clean up data before construction begins; delete data with duplicate time; verify and process coordinate data; clean up data with elevations lower than the foundation surface; clean up data that exceeds the lifting limit; convert the format of time data; and perform noise reduction on the data.

[0076] Step 3: Perform first-order difference calculation on the characteristic values ​​of the cable crane operation monitoring data and calculate the characteristic change rate, specifically:

[0077] The operation monitoring data of the cable crane includes information such as the cable crane station number, the cable crane coordinates at the time of data collection, the state duration, the collection time, and the cable crane weight. In this method, the cable crane weight (w) and the spatial coordinates (x, y) are referred to as the three-dimensional features of the cable crane. The change rate of each feature is calculated separately, that is, the ratio of the characteristic value to the time difference between two adjacent data. The weight change rate and trend state (the trend state is judged according to the positive or negative change rate. For example, when the weight change rate is negative, it means that the weight is decreasing, and when it is positive, it means that the weight is increasing), the displacement change rate and trend state, and the lifting depth change rate and trend state can be obtained. The calculation formulas are as follows: (1), (2) and (3) respectively. The present invention describes the operation state of the cable crane by calculating the change rate of the three-dimensional features of the cable crane.

[0078]

[0079] In formula (1), is the first-order difference value of the lifting weight, that is, the rate of change of the lifting weight; W n is the weight of the nth row, W n+1 is the lifting weight of the n+1th row, ΔW is the difference between the lifting weights of the previous and next two rows; t n is the acquisition time of the nth row, t n+1 is the acquisition time of the n+1th row, and Δt is the time difference between the previous and next two rows.

[0080] In formula (2), is the first-order difference value of the horizontal displacement, that is, the rate of change of the horizontal displacement; X n is the horizontal displacement of the nth row, Xn+1 is the horizontal displacement of the n+1th row, and ΔX is the difference in horizontal displacement between the previous and next rows.

[0081] In formula (3), is the first-order difference value of the vertical displacement, that is, the rate of change of the vertical displacement; Z n is the vertical displacement of the nth row, Z n+1 is the vertical displacement of the n+1th row, and ΔZ is the vertical displacement difference between the previous and next rows.

[0082] Step 4: Set the threshold to determine the trend status, specifically:

[0083] The three-dimensional eigenvalues ​​of the cable crane data are first-order differencing performed in chronological order. A threshold for change is set. When the difference value is within the threshold range, the state is considered stable. When it is greater than this range, the sign of the difference value is used to determine whether the state is increasing or decreasing. The measured data interval of the cable crane is 5 seconds. In this paper, the load threshold is set to 1 t, the X threshold is set to 2 meters, and the Z threshold is set to 1 meter. Each trend state is numbered: stable state is numbered "0"; rising state is numbered "1"; descending state is numbered "-1"; when the cable crane load is 0, it is marked as "0"; and when the cable crane is unloaded, it is numbered "2".

[0084] Then, the aggregate state is averaged, specifically:

[0085] Therefore, single time points with the same state are aggregated according to the trend of the 3D feature change. After aggregation, the mean of the feature values ​​within the state interval is calculated. After the state is aggregated, the initial time of the state is retained. This is used to determine the trend of 3D feature changes between two adjacent states.

[0086] Step 5: Use the trend status in step 4 to divide the single cycle, specifically:

[0087] Based on the trend state identification results of the above-mentioned single time point, boundary identification is performed. It is necessary to number all data in ascending order. When the trend state appears continuously, its starting point and end point are identified. Taking the cable crane lifting concrete as an example, it is considered that the cable crane goes from the no-load state to the lifting weight rising stage, the stable stage, the lifting weight descending stage, the stable stage, the lifting weight descending stage, and the no-load state is a cycle of the cable crane lifting concrete, where the no-load state is the boundary. For example, the no-load state in the previous cycle is the end point of the previous cycle, but it is also the starting point of the next cycle. Because the cyclical nature of the cable crane's operation is essentially the end-to-end connection of single cycles, the starting point of a single cycle is also the end point of the previous cycle. Taking the starting time of trend state "2" (cable crane no-load state) as the starting point, the single cycle division of the cable crane's measured data is completed. After completing the division of the single cycle, it is necessary to further determine the mode to which the cycle belongs based on the characteristics of the pattern cycle.

[0088] Step 6: Identify the four working modes of the cable crane by matching the parameters in Table 1, specifically:

[0089] In pattern recognition, a set of samples with the same or similar working states is called a pattern category. The working states of the cable crane are initially divided into four types: concrete lifting mode, miscellaneous object lifting mode, maintenance mode, and shutdown mode. The basic operating mode structure of the cable crane is as follows: Figure 2 As shown in the figure, the cable crane operating state is composed of a series of basic operating states. The cable crane operating state has distinct three-dimensional characteristics, encompassing both spatial location and temporal sequence. The significantly different lifting load characteristics between different modes—that is, the range of lifting load values ​​differs significantly, as does the number of lifts—are used to identify the cable crane operating mode. This difference in lifting load characteristics and the number of lifts between different cable crane modes is used to identify the cable crane operating mode.

[0090] The concrete lifting mode is actually the main part of the cable crane's work. There are two lifting times in a single cycle, and the distribution of its activity space is sequential. The concrete is lifted from the feeding platform to the warehouse surface for pouring. The lifting weight value rises from 0 (at this time, the cable crane is in an empty state and is preparing to lift an object) to the sum of the weight of the lifting tank and the concrete during the lifting process of the platform. It remains unchanged during the forward transportation. After arriving at the warehouse surface, it is unloaded and unloaded to the empty tank and then begins to return. The weight of the empty tank remains unchanged during the return process. The lifting weight is reduced to 0 when the lifting tank returns to the feeding platform. The specific structural mode is as follows Figure 3 shown.

[0091] The mode of lifting miscellaneous objects is also an important part of the operation of the cable crane. This mode has fewer cycles and has a significant lifting feature compared to lifting concrete, and the lifting is only done once in a single cycle. A single cycle of the mode of lifting miscellaneous objects will go through the following tasks: the cable crane moves empty to the loading location; waits for loading at the target location; the operator hangs heavy objects on the cable crane hook; lifts and rises; lifts the heavy load to the target location; the cable crane fine-tunes its position on the warehouse surface; and unloads at the target location. According to the different loading locations and target locations, it can be further subdivided into: "flat-warehouse" lifting mode, "warehouse-flat" lifting mode, warehouse surface transfer mode, and platform transfer mode. The specific structural modes are as follows: Figure 4 shown.

[0092] Maintenance mode is a special mode where there is no load involved and only movement occurs in the X and Z directions. The cableway is typically maintained twice a day, at approximately 6:00 and 18:00, for a typical one-hour maintenance period. During this maintenance period, the cableway will make a no-load run from the left bank to the right bank.

[0093] A cable crane operating under long periods of high load may enter shutdown mode at any time, except for scheduled daily maintenance. During shutdown mode, the crane's 3D feature values ​​generally remain unchanged, but fine-tuning may occur pending command. If the 3D feature values ​​remain unchanged for more than five minutes and do not occur at the time and location typically associated with maintenance mode, the crane is considered to have entered shutdown mode. The parameters for single-cycle identification of the crane's operating state are shown in Table 1 below.

[0094] Table 1 Operation mode single cycle identification parameter table

[0095]

[0096] If multiple heavy-load lifts occur within a single cycle (possibly due to the staged lowering of the load, a common occurrence when casting special structures), identification based solely on the number of heavy-load lifts is difficult. Therefore, a further scoring of the heavy-load lift status based on the number of lifts is used to distinguish between concrete lifting patterns and sporadic object patterns, thereby improving the accuracy of the recognition model. Since the number of lifts for concrete and zero lifting by cable cranes is the same—one return trip—identification based on the number of lifts is difficult. Therefore, scoring is based on the lifting weight characteristics of the lift times. For concrete lifting, the return load from the silo to the platform is generally around 6.5 tons, and the X-status trend is negative, whereas zero lifting does not exhibit this characteristic. Based on this characteristic, cable crane zero lifting can be distinguished from concrete lifting. This method uses a scoring method based on the lifting weight characteristics of heavy-load lifts to distinguish between concrete lifting and zero lifting patterns. The scoring principle is to distinguish the lifting characteristics of heavy-load lifting, distinguishing the two phases of heavy-load lifting, and assigning scores to each phase, known as the status score, with different values ​​set at different levels. The sum of the outbound and return scores is used to create a single value, known as the mode score. The specific scoring rules are shown in Table 2 below.

[0097] In the table, there's no conflict between loads greater than 1 ton and loads between 6 and 8 ton. This is because when querying SQL Server, the 6-8 ton load identification is performed first, followed by the load identification greater than 1 ton, and the two are not repeated. The X differential trend represents the changing trend of the X mean value in the heavy-load lifting state. This is calculated after state aggregation. The positive or negative value indicates whether the heavy-load lifting state is a forward or return trip.

[0098] Table 2 Single-loop recognition scoring table

[0099]

[0100] The cable crane's operating status is identified by the sum of the state scores within a single cycle, i.e., the mode. For the "sporadic lifting" mode, if the number of lifts identified is not 1 during a single lift, the forward state is assigned a score of 1, and the return state is assigned a score of 10. The total score can only be between 0 and 9 or divisible by 10. Any remaining score indicates the concrete lifting mode. The correspondence between the score and the operating mode is shown in Table 3 below.

[0101] Table 3 Correspondence between operation mode scores

[0102]

[0103] After all parameters and scoring rules are determined, the recognition process for a single cycle of the running mode is:

[0104] 1) Use the starting point of the cable car's no-load section to divide the operating mode into single cycles.

[0105] 2) If the no-load period exceeds 1 hour and is within the maintenance time range, it is identified as maintenance mode; if the other no-load periods exceed 5 minutes, it is identified as shutdown mode.

[0106] 3) Use SQL statements to statistically query trend status of "0" and duration greater than 25s, record the number of heavy-load lifting times, and then identify the lifting patterns of concrete and miscellaneous objects based on the scoring rules.

[0107] Step 7: Based on the recognition results of step 6, match the three-dimensional features and identify the 14 cable crane working states shown below, specifically:

[0108] After dividing the measured data into single cycles and determining the operating mode, it is necessary to further identify the working status within the single cycle. Each piece of data carries relevant information such as trend status, lifting weight, coordinates, etc., and the working status is identified within the cycle based on the three-dimensional feature information of the data. The specific process is as follows: Figure 5 The specific identification method is as follows:

[0109] (1) Waiting for loading: This stage is the initial state of a single cycle, with the cable crane hook stationary on the platform or warehouse surface. All three eigenvalues ​​are stable, with the same trend. The load is between 0 and 1 t. If the X value is within the platform range and the elevation is at the platform elevation, the platform is waiting for loading. If the X value is within the dam range and the elevation is below the transport elevation, the warehouse surface is waiting for loading. Therefore, the SQL keyword casewhen, combined with the trend state and the eigenvalue range, can be used to identify each.

[0110] (2) Lifting and Ascending: During the entire lifting process, the main characteristic is that the lifting weight value rises from no load to the lifting weight value of the target object. If the lifting weight is fully loaded with concrete, the lifting weight is about 28 tons. This process rarely pauses, and the W trend state is 1, which is the only one that exists in all states. The difference is that the lifting and ascending in the concrete lifting mode occurs within the platform range, while the "warehouse-level" lifting mode and transfer mode occur within the dam body range. Therefore, the rate of change of the W value is identified as rising, and the lifting position is determined in combination with the X value range.

[0111] (3) Transportation preparation: This state is a transitional state with a short duration. After the cable car lifts the heavy object, it stays still in the air for a period of time. The trend state of the three eigenvalues ​​is 0. The W value is heavy load lifting, and the elevation is within the transportation elevation range. The X value can determine whether it is above the platform or above the warehouse surface.

[0112] (4) Horizontal transportation: After transportation preparation, the cable car starts horizontal heavy-load transportation. The lifting weight value and Z value trend state are 0, the X value trend state is 1, the elevation is within the transportation elevation range, and the W value is heavy-load lifting weight.

[0113] (5) Descending into the warehouse: The cable crane runs above the warehouse surface and descends into the warehouse surface. The trend state of the lifting weight value and the X value is 0, the trend state of the Z value is -1, the W value is heavy load lifting, and the X value is within the range of the dam body.

[0114] (6) Warehouse surface alignment: After the cable crane bucket enters the warehouse under the command of the commander, the unloading point may need to be changed and slight adjustments will be made. Since the adjustment speed is slower than that during horizontal transportation, the X value trend state is still judged as 0 during identification, and the hoisting weight value and Z value trend state are also 0. The elevation is the warehouse surface elevation, the X value is within the dam body range, and the W value is the heavy load hoisting weight.

[0115] (7) Unloading: During the unloading process, the W value drops significantly, the trend state is -1, and the X value trend state is 0. During unloading, the hanging tank will rise slightly and the speed is slow. The Z value is still considered as 0. When the pouring elevation is close to the elevation of the feeding platform, it will cause a misjudgment of the descending return state. The Z value range should not be judged by the elevation, but can only be limited by the elevation below the transportation elevation. At the same time, the X value range is used to distinguish the descending return state.

[0116] (8) Preparation for leaving the warehouse: Before the cable crane hook leaves the warehouse, the trend state of the three characteristic values ​​is 0. If the concrete is lifted, the W value of the empty tank is between 6 and 8 tons, and the W value of the zero lifting mode is heavy load. In addition, the X value is within the range of the dam body and the Z value is lower than the transportation elevation range. This feature is unique and not easy to misjudge.

[0117] (9) Ascending out of the warehouse: During the ascending process of the cable crane hook, the Z value trend state is 1. The Z value increases only in the hoisting ascending state, but the W value also increases in the hoisting ascending state. Therefore, combined with the W and X value trend states of 0, the ascending out of the warehouse state can be directly identified.

[0118] (10) Empty tank return: During the empty tank return process, the X value trend state is -1. Compared with the zero hanging mode, the empty tank return has the characteristics of lifting weight. The W value of the zero hanging mode return is empty. Combined with the W and Z value trend states of 0, the empty tank return state can be directly identified.

[0119] (11) Platform alignment: After the empty tank returns, the cable crane hook is aligned on the platform. The characteristics of this process are that the trend state of the three characteristic values ​​is 0, the X value is within the platform range, the elevation value is greater than the platform elevation, and the empty tank in the mixed lifting mode is between 6 and 8 tons, which is easy to identify; the lifting range of the zero lifting mode is larger, so the W value is not constrained.

[0120] (12) Lowering and returning: The lowering and returning process includes the lowering and unloading of the cable crane hook. Since the platform unloading duration is short, it is incorporated into the lowering and returning state. In this process, the X value trend state is 0, which is within the platform range; the Z value trend state is -1 at first, and then changes to 0 when unloading. At this time, the W value trend state is -1; the empty tank lifting weight in the mixed lifting mode is between 6 and 8 tons; the lifting weight range in the zero lifting mode is larger, so the W value is not constrained.

[0121] (13) Maintenance state: The maintenance state has obvious hoisting characteristics and a long duration, which can be distinguished from the waiting state of the mixed hoisting mode and the zero hoisting mode. Therefore, the method to identify the maintenance state is to use a hoisting value between 0 and 1t, a trend state of 2, and an occurrence time within the maintenance period. The SQL keyword datepart can be used to obtain the hour integer value at any time, and the duration threshold is set to 1h.

[0122] (14) Stoppage state: The lifting characteristics of the stoppage state and the maintenance state are the same. The key to distinguishing them is the duration and the time of occurrence. The identification method is to set the threshold to 5 minutes when the no-load duration exceeds the upper limit of the waiting time for the mixed and zero-load modes outside the maintenance period.

[0123] The technical solution provided by the present invention can effectively identify the working status of the cable crane based on massive monitoring data. The method of the present invention is verified using test data. The original cable crane operation data extracted from the database on January 1, 2020 for cable crane 1# is as follows: Figure 6 As shown. The single time points with the same state are aggregated according to the trend state of the weight change, and the mean of the characteristic values ​​in the state interval is calculated respectively. After the state is aggregated, the initial moment of the state is retained, as shown below Figure 7After the cable machine single cycle identification process, some results are as follows Figure 8 Taking the data of 5# cable crane on May 1, 2019 as an example, the working status recognition results are shown in Figure 9 .

Claims

1. A method for identifying the working status of a cable crane based on massive monitoring data, characterized by: The following steps are involved: 1) Obtain cable crane operation monitoring data; 2) Preprocessing the cable crane operation monitoring data obtained in step 1); 3) Calculate the rate of change of the three-dimensional characteristics of the cable crane based on the pre-processed cable crane operation monitoring data. The specific implementation method is: 3.1) Obtaining pre-processed cable crane operation monitoring data, wherein the pre-processed cable crane operation monitoring data includes the cable crane station number, cable crane coordinates at the time of data collection, state duration, collection time, and cable crane hoisting weight; 3.2) extracting the three-dimensional features of the cable crane from the pre-processed cable crane operation monitoring data; the three-dimensional features of the cable crane include the cable crane hoisting weight and spatial coordinates; 3.3) Calculate the change rates of the three-dimensional features of the cable crane in step 3.2) respectively; the change rates of the three-dimensional features of the cable crane include the rate of change of the lifting weight, the rate of change of the horizontal displacement, and the rate of change of the lifting depth of the cable crane. The calculation method of the lifting weight change rate is: ， The horizontal displacement change rate is calculated as follows: ， The calculation method of the cable crane lifting depth change rate is: ， in: is the first-order difference value of the lifting weight, that is, the rate of change of the lifting weight; W n is the weight of the nth row; W n+1 It is the weight of the n+1th row; ΔW is the difference in lifting weight between the front and rear rows; is the first-order difference of the horizontal displacement, that is, the rate of change of the horizontal displacement; X n is the horizontal displacement of the nth row; X n+1 is the horizontal displacement of the n+1th row; ΔX is the horizontal displacement difference between the two rows; is the first-order difference of vertical displacement, that is, the rate of change of vertical displacement; Z n is the vertical displacement of the nth row; Z n+1 is the vertical displacement of row n+1; ΔZ is the vertical displacement difference between the two rows; t n is the acquisition time of the nth row; t n+1 is the acquisition time of row n+1; Δt is the time difference between the two rows; 4) Determine the trend state of the cable crane based on the rate of change of the three-dimensional characteristics of the cable crane obtained in step 3), specifically by: 4.1) Setting the change threshold; 4.2) Determine whether the rate of change of the three-dimensional characteristics of the cable crane is within the change threshold set in step 4.1). If so, the cable crane is considered to be in a stable state; if not, proceed to step 4.3); 4.3) Determining whether the cable crane is in an ascending or descending state based on the positive or negative change rate of the cable crane's three-dimensional characteristics; when the change rate of the cable crane's three-dimensional characteristics is positive, it indicates that the cable crane is in an ascending state; when the change rate of the cable crane's three-dimensional characteristics is negative, it indicates that the cable crane is in a descending state; 4.4) Aggregate single time points with the same state according to the trend of 3D feature changes. After aggregation, calculate the mean of the feature values ​​within the state interval. After aggregation, retain the initial time of the state; this is used to determine the trend of 3D feature changes between two adjacent states. 5) Using the trend state of the cable crane determined in step 4), the measured data of the cable crane is divided into single cycles. The specific implementation method is: 5.1) Perform boundary identification based on the trend state identification results at the above single time point; 5.2) Numbering the results of step 5.1) in ascending order; 5.3) When the trend state of the cable car appears continuously, identify its starting point and end point; 5.4) Taking the starting time of the cable crane’s trend state as the starting point, complete the single-cycle division of the cable crane’s measured data; 6) matching the identification parameters of the single cycle in step 5), and identifying the four working modes of the cable crane according to the identification parameters; wherein the identification parameters include the number of heavy load hoisting times, hoisting weight characteristics, and time characteristics; The four working modes of the cable crane are: concrete lifting mode, miscellaneous object lifting mode, maintenance mode and shutdown mode; The identification parameters of the concrete lifting mode are: the number of heavy-load lifting is 2 times; the lifting weight is not higher than 28 tons on the forward trip and not lower than 6.5 tons on the return trip; The identification parameters of the lifting mode of sporadic objects are: the number of heavy-load lifting is 1; the lifting weight characteristic is 1 to 30 tons; The identification parameters of the maintenance mode are: the number of heavy-load lifting is 0 times; the lifting weight characteristic is 0 to 1 t; the time characteristic is 1 hour; The identification parameters of the shutdown mode are: the number of heavy-load lifting is 0 times; the lifting weight characteristic is 0 to 1 t; the time characteristic is ≥ 5 minutes; If multiple heavy-load lifting occurs in one cycle, it is difficult to identify based on the number of heavy-load liftings alone. Based on the relationship between the number of liftings, the heavy-load lifting status is further scored to distinguish between the concrete lifting mode and the sporadic object lifting mode; The scoring rules are as follows: If the lifting weight of a heavy-load lift in a single cycle is 6 to 8 tons and its X-difference trend is negative, the heavy-load lift mode is classified as 10; If the lifting weight of a heavy-load lift in a single cycle is 6 to 8 tons and its X-difference trend is positive, the heavy-load lift mode is classified as 1; If the lifting value of a heavy load in a single cycle is greater than 1t and its X difference trend is positive, the heavy load lifting mode is classified as 1; If the lifting value of a heavy load in a single cycle is greater than 1t and its X difference trend is negative, the heavy load lifting mode is classified as 1; X differential trend, positive indicates that the heavy load lifting state is forward, negative indicates that the heavy load lifting state is return; Then, the mode scores of all heavy load lifting in a single cycle are summed up; If the sum of the scores is 1 to 9, the single cycle is determined to be a mode for lifting sporadic objects; If the sum of the scores is a multiple of 10, the single cycle is determined to be a mode for lifting scattered objects; If the sum of the scores is the remaining scores, the single cycle is determined to be the concrete lifting mode; 7) Based on the recognition results of step 6), the three-dimensional features are matched to identify the working status of the cable crane within a single cycle; the cable crane working status includes waiting for loading, lifting and ascending, transportation preparation, horizontal transportation, descending into the warehouse, warehouse surface alignment, unloading, warehouse out of the warehouse, ascending out of the warehouse, empty tank return, platform alignment, descending back to the original position, maintenance status, and shutdown status.

2. The cable crane working status identification method based on massive monitoring data according to claim 1 is characterized in that: The specific implementation method of step 1) is: using SQL statements to extract the required data from the database of the cable crane information monitoring system.

3. The cable crane working status identification method based on massive monitoring data according to claim 1 is characterized in that: The pre-processing in step 2) is to delete the time duplicate data, delete the erroneous data, verify and process the coordinate data, convert the format of the time data and perform noise reduction on the cable crane operation monitoring data.

Images