A Construction Data Processing Method and System Based on Big Data

Abstract

The present invention relates to the technical field of data processing, and particularly relates to a construction data processing method and system based on big data. The method includes: synchronously collecting the load data of a drilling machine and the current data of an electric motor, setting historical load data for each load data to be predicted, calculating the prominence degree of the historical load data, and combining the change trend of the current data to optimize the prominence degree of the historical load data to obtain the credibility of the historical load data. Based on the credibility of all historical load data, accurately determine the reference value of the load data to be predicted, use this reference value to reasonably adjust the prediction range of the load data to be predicted, and then accurately calculate the predicted value of the load data to be predicted. Judge the abnormal load data according to the difference between the predicted value and the actual value, improve the accuracy of monitoring the operating state and fault prediction of the drilling machine, and thus achieve more accurate and reliable processing of the construction process data.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing. Specifically, it relates to a method and system for processing construction data based on big data. Background Art

[0002] During the construction process of a building, which covers various stages such as project initiation, design, procurement, construction, and acceptance, a vast amount of data is generated. The sources of this data include information collected by sensors, records in construction logs, details of material procurement, and labor distribution. Drilling operations, as a fundamental and crucial link in the construction process, and drilling machines, as the core heavy equipment among them, whether their stable operation is directly related to the progress of the entire construction project, cost control, and safety assurance. Therefore, accurately analyzing and processing the relevant data of the drilling machine during the construction process, carefully analyzing its operating state, and predicting potential safety hazards in advance so as to implement maintenance measures in a timely manner are of great significance for ensuring the normal and orderly progress of the construction.

[0003] In the existing technical scope of equipment fault prediction, usually, the operating data of the equipment is first collected, and then the HTFE prediction algorithm is used. This prediction algorithm predicts the future operating data of the equipment based on the past historical data of the equipment's operation, so as to determine whether there are potential faults in the equipment. For example, the Chinese patent application document with the publication number CN119146267A discloses a method and system for controlling a water supply valve based on fault monitoring. It collects the pressure and valve diameter data of the valve, respectively obtains the pressure sequence and valve diameter sequence, then predicts these sequences at historical moments based on the HTFE algorithm, obtains the predicted values of the pressure and valve diameter at the current moment, and corrects the sum of the differences between the predicted values and their respective actual values using the reliability of the predicted values at the current moment, thereby obtaining the fault degree at the current moment.

[0004] However, there is a defect in the HTFE prediction algorithm used in the above technical solution. Specifically, this prediction algorithm calculates the prediction range of the current data based on the prediction range of the previous data and the actual value of the previous data, and then determines the predicted value of the current data. In the actual construction data environment, especially in the acquisition link of the drilling machine load data, due to the inevitable strong mechanical vibration during the drilling operation, this vibration is very likely to interfere with the signal acquisition of the load sensor, and then lead to the generation of load noise data. Once the noise data is mixed in, the actual value of the previous load data is very likely to be contaminated and transformed into noise data. In this case, the predicted value of the current load data calculated by the HTFE algorithm will deviate from the real situation, which will further cause deviations in the prediction of the overload fault of the drilling machine, thus affecting the subsequent abnormal judgment of the load data and the prediction of the overload fault of the drilling machine, and unable to meet the requirements for data accuracy and reliability during the construction process. Summary of the Invention

[0005] To solve the problem that in the current construction environment, the strong mechanical vibration generated during the drilling operation often interferes with the signal acquisition of the load sensor, resulting in the load data of the drilling machine being mixed with noise during the acquisition, affecting the prediction accuracy of the traditional HTFE algorithm for the operation data of the drilling machine, and then leading to the inability to accurately predict the potential faults of the drilling machine, affecting the construction progress and safety, the present invention proposes a construction data processing method based on big data, including:

[0006] Collect the load data of the drilling machine during the construction process and the current data of the electric motor at the same time. Take each load data as the load data to be predicted, and use the HTFE prediction algorithm to obtain the prediction range of each load data to be predicted;

[0007] Set multiple historical load data for each load data to be predicted, calculate the prominence of each historical load data, and evaluate the credibility of each historical load data according to the prominence of each historical load data, the load data within the adjacent data segment of the historical load data, and the change trend of the corresponding current data within the adjacent data segment;

[0008] Determine the reference value for adjusting the prediction range of the load data to be predicted based on the credibility of all historical load data of the load data to be predicted, use the reference value to adjust the prediction range of the load data to be predicted, calculate the predicted value of the load data to be predicted according to the adjusted prediction range of the load data to be predicted, and determine the abnormal load data according to the difference between the predicted value and the actual value of each load data, so as to realize the processing of the data of the drilling machine during the construction process.

[0009] The above technical solution obtains key information reflecting the working state by collecting the load data of the drilling machine during the construction process. The motor current data is closely related to it and can reflect the operating conditions of the drilling machine from the side. Taking the load data as the object to be predicted, the HTFE algorithm is used to initially determine its prediction range, which is the basis for subsequent more accurate prediction and analysis. Further, considering the importance of historical data for the current prediction, not all historical data has the same reliability. By calculating the prominence degree and combining the load and current data change trends of adjacent data segments to evaluate the credibility, the quality of different historical data can be distinguished. The prominence degree can reflect the abnormal degree of a certain historical data relative to other data, while the change trend reflects the rationality and stability of the data in the time series. The combination of the two can screen out relatively reliable historical data and exclude the interference of unreliable historical data caused by factors such as noise and abnormal working conditions on the prediction. Further, after determining the credibility of the historical data, a reference value for adjusting the prediction range is generated based on this, and the prediction range initially obtained by the HTFE algorithm is corrected to make the prediction range more in line with the actual situation. Then, more accurate prediction values are calculated according to the adjusted prediction range, and finally, by comparing the differences between the prediction values and the actual values, abnormal load data is identified, so as to effectively process the construction data and achieve the purpose of accurately monitoring the operating state of the drilling machine and discovering potential problems.

[0010] Preferably, the credibility of each historical load data satisfies the following relational expression:

[0011] ;

[0012] In the formula, is the credibility of the th historical load data of the th load data to be predicted, is the prominence degree of the th historical load data of the th load data to be predicted, is the value of the current data corresponding to the collection time of the th historical load data of the th load data to be predicted, is the mean value of the adjacent data segment of the th historical load data of the th load data to be predicted, is the value of the th historical load data of the th load data to be predicted, is the th historical load data of the The mean value of the current data within the adjacent data segment of the historical load data, is the absolute value symbol, and is the natural exponential function.

[0013] The various parameters in the above technical solution reflect the internal relationships between different data. The mean values of the current data and the load data within the adjacent data segment and their relative relationships are used to measure the credibility of the historical load data, which reflects the physical relationship between the load and the current during the operation of the drilling machine, that is, the current change of the motor is often related to the load change of the drilling machine. Therefore, this data association logic in actual operation is introduced in the credibility evaluation, making the calculation of credibility more in line with the actual working conditions.

[0014] Preferably, the calculation formula for the prominence degree of each historical load data is:

[0015] ;

[0016] In the formula, is the prominence degree of the th historical load data of the th load data to be predicted, is the prediction error of the th historical load data of the th load data to be predicted, is the value of the th historical load data of the th load data to be predicted, is the mean value of the adjacent data segment of the th historical load data of the th load data to be predicted, is a hyperparameter, is the change amount of the th historical load data of the th load data to be predicted. The change amount of the th historical load data is equal to the absolute value of the difference between the th historical load data and the th historical load data.

[0017] The above technical solution not only focuses on the static value of the historical load data but also considers its dynamic change characteristics. In actual construction, the load of the drilling machine changes with the change of operating conditions. Capturing this dynamic characteristic helps to more accurately identify the data that shows abnormalities during the change process, so as to better adapt to the complex working conditions and data fluctuations during the construction process, making the evaluation of the historical load data more in line with the actual operation situation.

[0018] Preferably, the reference value for adjusting the prediction range of the load data to be predicted satisfies the following relational expression:

[0019] ；

[0020] In the formula, is the reference value for adjusting the prediction range of the th load data to be predicted, is the total number of historical load data of the th load data to be predicted, is the credibility of the th historical load data of the th load data to be predicted, is the sum of the credibilities of all historical load data of the th load data to be predicted, is the actual value of the th historical load data of the th load data to be predicted.

[0021] The above technical solution makes the historical load data with high credibility have a greater influence when determining the reference value, while the historical load data with low credibility has a relatively smaller influence. This weight allocation mechanism based on credibility can effectively utilize reliable historical data, reduce the interference of unreliable data on the adjustment of the prediction range, and make the reference value more able to reflect the real situation.

[0022] Preferably, the acquisition method of the adjacent data segment of the historical load data is as follows: Combine each historical load data with multiple adjacent load data before its acquisition time to form the adjacent data segment of the historical load data.

[0023] Preferably, the method for adjusting the prediction range of the load data to be predicted using the reference value is:

[0024] If the upper limit value of the prediction range of the th load data to be predicted is less than , let the upper limit value of the prediction range of the th load data to be predicted be equal to ; if the lower limit value of the prediction range of the th load data to be predicted is greater than , let the lower limit value of the prediction range of the th load data to be predicted be equal to .

[0025] The adjustment method of the above technical solution constrains the upper and lower limits of the prediction range based on the reference value to ensure that the prediction range does not deviate from the reasonable interval determined based on the characteristics of historical data. If the upper limit value of the prediction range is too small or the lower limit value is too large, which does not conform to the actual situation reflected by the historical data compared with the reference value, by adjusting it to the reference value, the prediction range can be made more in line with the internal laws and trends of the data, avoiding unreasonable prediction ranges caused by factors such as the limitations of the prediction algorithm itself or data noise.

[0026] Preferably, the method for calculating the predicted value of the load data to be predicted according to the adjusted prediction range of the load data to be predicted is as follows:

[0027]

[0028] In the formula, is the predicted value of the th load data to be predicted, is the upper limit value of the adjusted prediction range of the th load data to be predicted, is the lower limit value of the adjusted prediction range of the th load data to be predicted, , , are all parameters of the HTFE prediction algorithm.

[0029] Preferably, the method for determining abnormal load data according to the difference between the predicted value and the actual value of each load data is as follows:

[0030] Obtain the normalized value of the difference between the predicted value and the actual value of each load data. If the normalized value is greater than the preset difference threshold, the load data is abnormal load data.

[0031] Preferably, after determining the abnormal load data according to the difference between the predicted value and the actual value of each load data, the following operations are also performed:

[0032] Continuously monitor. If N abnormal load data appear within a certain minute, issue a warning to notify relevant personnel to maintain the drilling machine, where N is a preset value.

[0033] The present invention also provides a construction data processing system based on big data. The construction data processing system includes a memory and a processor. A computer program is stored on the memory, and the processor executes the computer program to implement the steps of any one of the construction data processing methods.

[0034] The present invention has the following effects:

[0035] When determining the prediction range of each load data to be predicted, the present invention introduces multiple historical load data of the load data to be predicted, takes into account the characteristics and change trends of the historical load data, combines the change trends of the corresponding current data, accurately evaluates the credibility of the historical load data, and performs weighted average calculation on the multiple historical load data to predict the range, which can effectively identify and eliminate the interference of noise data, so as to ensure that the prediction range calculated by the HTFE algorithm more accurately reflects the actual load change trend of the drilling machine, thereby improving the accuracy of the prediction of the operation data of the drilling machine, making the potential fault prediction result of the drilling machine more reliable, and ensuring the normal progress and safety of the construction to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] By referring to the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown in an exemplary rather than restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0037] Figure 1 is a schematic flowchart of the method of the present invention;

[0038] Figure 2 is a schematic flowchart of the method of step S2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0040] The following will describe in detail the specific embodiments of the present invention with reference to the accompanying drawings.

[0041] Refer to Figure 1 , a construction data processing method based on big data provided by the present invention includes the steps - Step :

[0042] S1: Collect the load data of the drilling machine and the current data of the motor during the construction process.

[0043] In one embodiment, by installing a high-precision force sensor at the drill bit part of the drilling machine, accurate real-time acquisition of the load data of the drilling machine is implemented. At the same time, a current sensor is used to synchronously monitor the current data of the motor in real time. To ensure the effectiveness and relevance of the data, it is ensured that the acquisition of the two types of data is carried out within a strictly identical time period and is precisely corresponding on the time axis. The acquisition duration is set to one hour, and the acquisition frequency is fixed at once per second, so as to obtain the load of the drilling machine and the motor current data with high resolution and matching time series, providing a solid data basis for subsequent analysis and processing.

[0044] S2: Analyze the numerical change characteristics of the load data of the drilling machine and the current data of the motor, optimize the way of adjusting the prediction range of the load data to be predicted by the HTFE model, and obtain the adjusted prediction range of the load data to be predicted.

[0045] In one embodiment, the method for optimizing the way of adjusting the prediction range of the load data to be predicted by the HTFE model is as Figure 2 shown:

[0046] S21: Analyze the numerical characteristics of each historical load data of the load data to be predicted, and obtain the prominence degree of each historical load data of the load data to be predicted.

[0047] Specifically, taking any one load data as the load data to be predicted, the load data before the load data to be predicted are used as all the historical load data of the load data to be predicted, (empirical value). Since when collecting the load data of the drilling machine, noise data often show obvious characteristics, that is, their values will fluctuate rapidly in a short period of time, showing extremely high or extremely low values significantly different from the normal operation data. Therefore, when analyzing the prominence degree of each historical load data, the following key factors are comprehensively considered: First, the magnitude of the error generated when each historical load data is predicted using the HTFE prediction algorithm. A larger prediction error indicates that the historical load data deviates more from the value expected based on the historical trend. Second, compare the value of this historical load data with the mean value of its adjacent data segment. If the historical load data deviates more from this mean value, the particularity of this historical load data is more obvious. Finally, examine the numerical change amount of this historical load data and the historical load data at the previous moment. If the change amount is large, this may also imply that this historical load data is affected by abnormal factors such as noise, resulting in a sudden change in the value. Considering these three aspects of factors, the prominence degree of each historical load data is determined.

[0048] In one embodiment, for each piece of historical load data, the adjacent data segment is obtained as follows: Combine this historical load data with the 50 (empirical value) load data immediately preceding its acquisition time to form its adjacent data segment. However, if the number of load data before the acquisition time of a certain piece of historical load data is less than 50, then the load data within the previous two minutes is selected to make up the difference, so as to ensure that the adjacent data segment of each piece of historical load data can be completely constructed, thereby providing a stable and reliable data basis for subsequent data analysis based on the adjacent data segment, and ensuring the coherence and accuracy of the entire construction data processing process.

[0049] In one embodiment, the prominence of each piece of historical load data is calculated according to the following formula:

[0050]

[0051] In this formula, is the prominence of the th historical load data of the th load data to be predicted, is the prediction error of the th historical load data of the th load data to be predicted, is the value of the th historical load data of the th load data to be predicted, is the mean value of the adjacent data segment of the th historical load data of the th load data to be predicted, is a hyperparameter, (empirical value), The existence of is to prevent the situation where is 0. is the change amount of the th historical load data of the th load data to be predicted. The change amount of the th historical load data is equal to the absolute value of the difference between the th historical load data and the

[0052] In this formula, intuitively reflects the abnormal fluctuation of the value of the th historical load data of the th load data to be predicted. The larger is, the greater the possibility that the th historical load data has an abnormal value fluctuation, indicating that the The more likely a historical load data is to be potential noise data, the greater its prominence. It measures the deviation of the historical load data from the mean of the adjacent data segments. A larger deviation indicates that the historical load data behaves abnormally in the local range. The greater the deviation from the mean, the higher the probability of becoming noise data and the greater the prominence. This consideration based on the data distribution characteristics makes the anomaly judgment more scientific. Specifically, The larger the value of the historical load data, the more it deviates from the mean of the adjacent data segments of the historical load data, which means the historical load data is more likely to be potential noise data, and thus its prominence is greater. focuses on the change amount between each historical load data and the load data at the previous moment. During the construction process, under normal working conditions, the change of the load data is relatively stable, while noise interference will cause sudden changes in the load data. By The larger the value of the change amount between the historical load data and the load data collected at the previous moment, the more the historical load data deviates from the mean of the adjacent data segments of the historical load data, and the more likely the

[0053] historical load data is to be potential noise data, and its prominence will also be greater.

[0054] S22: Optimize the prominence of each historical load data according to the prominence of each historical load data, the load data within the adjacent data segments of the historical load data, and the change trend of the current data corresponding to the adjacent data segments, to obtain the credibility of each historical load data.

[0055] After a series of previous steps of analysis, although the prominence of each historical load data for each load data to be predicted is obtained, this prominence is only determined based on the numerical performance of each historical load data itself, and it is impossible to distinguish whether it is caused by noise data interference or large fluctuations in load data.

[0056] This step takes into account that drill rigs are usually driven by electric motors, and there is a proportional relationship between the load of the electric motor and its current or power consumption. Specifically, during drilling operations, when the formation resistance increases and the load of the drill rig increases, the electric motor has to output more power to cope, and at this time the current will increase accordingly, which reflects the close and practical logical connection between the two. Therefore, on this basis, further analyze the numerical performance of each historical load data and its corresponding current data. By observing the correlation between the two, if the positive correlation between a certain historical load data and its corresponding current data is worse, it means that this historical load data is more likely to be an anomaly caused by noise interference rather than a real load change under normal construction conditions. Then, the credibility assigned to this historical load data should be lower.

[0057] In this way, optimize the prominence obtained solely based on the historical load data itself before, and obtain the credibility that can better reflect the true and reliable degree of the historical load data, so as to more accurately utilize these historical data in subsequent operations such as predicting load data and judging anomalies, and reduce problems such as inaccurate prediction caused by misjudging the nature of the data (mistaking noise data for real abnormal data or vice versa).

[0058] Generally speaking, it is to further analyze and optimize on the basis of the prominence of each historical load data in combination with the current data of the electric motor, so as to more accurately judge the reliability of each historical load data.

[0059] In one embodiment, the credibility of each historical load data satisfies the following relational expression:

[0060]

[0061] In the formula, is the credibility of the th historical load data of the th load data to be predicted, is the prominence of the th historical load data of the th load data to be predicted, is the value of the current data corresponding to the acquisition moment of the th historical load data of the th load data to be predicted, is the The mean of the adjacent data segment of the th historical load data for the load data to be predicted, is the th value of the th historical load data for the load data to be predicted, is the th mean of the current data corresponding to the adjacent data segment (based on the acquisition time) of the th historical load data for the load data to be predicted, is the absolute value symbol, is the natural exponential function.

[0062] In this formula, the larger it is, the greater the possibility that the th historical load data for the load data to be predicted is noise data rather than real data, and then its corresponding credibility is smaller. quantifies the numerical ratio between the adjacent data segment of the th historical load data and the corresponding current data segment of this adjacent data segment, represents the reference value of the th historical load data. The greater the absolute value of the difference between this reference value and , it can indicate that the positive correlation between the th historical load data and the numerical performance of its corresponding current data is worse, indicating that the th historical load data is more likely to be noise data rather than real data. Then the th historical load data will have a lower corresponding credibility. In summary, by introducing the current data and the mean of the historical load data and its adjacent data segment, a mechanism reflecting the internal relationship between the load data and the current data is constructed. In actual construction, as mentioned above, when the motor drives the drilling machine, there is a close positive correlation between the load and the current. Based on this actual working condition, this formula reflects this association in a mathematical form, that is, by calculating the absolute value of the difference between the reference value and the value of the historical load data itself to measure the correlation between the two, making the calculation of the credibility more in line with the operation principle of the drilling machine and the change law of the actual data, screening out possible noise data interference, and then being able to improve the accuracy of the subsequent prediction of the drilling machine load data and more accurately reflect the actual load change situation of the drilling machine.

[0063] S23: Determine a reference value for adjusting the prediction range of the load data to be predicted based on the credibility of all historical load data of the load data to be predicted.

[0064] S23: Determine a reference value for adjusting the prediction range of the load data to be predicted based on the credibility of all historical load data of the load data to be predicted.

[0065] After obtaining the credibility of each historical load data for each load data to be predicted, in this step, a reference value for adjusting the prediction range of the load data to be predicted is calculated by weighted average based on this index and the actual value of each historical load data, and the prediction range of the load data to be predicted is accurately adjusted according to this reference value in the subsequent steps. If the credibility of a historical load data is greater, it means that the possibility of this historical load data point belonging to noise data is smaller, then the actual value of this historical load data should occupy a greater weight when calculating this reference value.

[0066] In one embodiment, the reference value for adjusting the prediction range of the load data to be predicted satisfies the following relational expression:

[0067]

[0068] In the formula, is the reference value for adjusting the prediction range of the th load data to be predicted, is the total number of historical load data of the th load data to be predicted, and the empirical value means that 10 historical load data are selected for each load data to be predicted, is the credibility of the th historical load data of the th load data to be predicted, is the sum of the credibilities of all historical load data of the th load data to be predicted, is the actual value of the th historical load data of the th load data to be predicted.

[0069] This formula calculates the reference value by combining the credibility and the actual value of each historical load data, reflecting an idea of weighted average based on credibility. This means that instead of simply taking the conventional average of the actual values of all historical load data, corresponding weights are assigned according to the reliability of each historical load data, so that more credible data can play a greater role in determining the reference value, which is more in line with the logic of focusing on the use of reliable data in actual data processing and scientifically integrates the information contained in historical load data.

[0070] To sum up, the significance of the reference value for adjusting the prediction range of the load data to be predicted lies in:

[0071] Improve the rationality of prediction range adjustment: The reference value calculated by this method of weighted average based on credibility provides a scientific and reasonable basis for the subsequent adjustment of the prediction range of the load data to be predicted.

[0072] Enhance the resistance to noise data: When facing possible noise data mixed in, since the data with low credibility has a small weight in calculating the reference value, its influence on the reference value is effectively suppressed, and the reference value will not deviate from the normal range due to the interference of individual noise data. In this way, when adjusting the prediction range using the reference value, the adverse effects brought by the noise data can be reduced.

[0073] S24: Adjust the prediction range of the load data to be predicted through the above reference value.

[0074] In one embodiment, the prediction range of the load data to be predicted is obtained by using the HTFE prediction algorithm:

[0075] First, perform the initialization operation:

[0076] First, obtain the prediction range of the load data to be predicted as the initial condition. The specific method is as follows:

[0077] Take the 20 load data before the load data to be predicted as the load data sequence of the first load data to be predicted. Take the mean value of this load data sequence as the predicted value of the first load data to be predicted. Take the maximum value and the minimum value in this load data sequence as the upper limit value and the lower limit value of the maximum prediction range of the first load data to be predicted respectively. Then, the prediction range of the first load data to be predicted is composed of this upper limit value and lower limit value.

[0078] Next, for the load data to be predicted , the HTFE prediction algorithm obtains the prediction range of the load data to be predicted according to the actual value of the previous load data of the load data to be predicted , the prediction error of the previous load data and the prediction range of the previous load data. The method is as follows:

[0079] Calculate the initial predicted value of the load data to be predicted :

[0080] ;

[0081] In this formula, is the error factor of the HTFE prediction algorithm, .

[0082] Then, the prediction range of the load data to be predicted is:​

[0083] ;

[0084] ;

[0085] In the above formula, and are respectively the upper limit value and the lower limit value of the prediction range of the -th load data to be predicted, and are respectively the upper limit value and the lower limit value of the prediction range of the -th load data to be predicted, is the influence factor of the prediction error range of the HTFE prediction algorithm, .

[0086] In one embodiment, the method for adjusting the prediction range of the traditional HTFE algorithm for the load data to be predicted is as follows: when the upper limit value of the prediction range of the load data to be predicted is less than the initial prediction value of the load data to be predicted, the upper limit value of the prediction range of the load data to be predicted is set equal to the initial prediction value of the load data to be predicted; when the lower limit value of the prediction range of the load data to be predicted is greater than the actual value of the previous load data, the lower limit value of the prediction range of the load data to be predicted is set equal to the initial prediction value of the load data to be predicted.

[0087] Specifically, it is to ensure that the predicted value of the load data to be predicted can cover the actual value of the previous load data , when is less than , let be equal to , when is greater than , let be equal to .

[0088] This method cannot guarantee whether the actual value of the previous load data is the true value, and is extremely vulnerable to noise, which may lead to inaccurate calculation of the prediction range of the load data to be predicted, and further affect the accuracy of calculating the predicted value of the load data to be predicted.

[0089] In one embodiment, the method for adjusting the prediction range of this step for the load data to be predicted is as follows:

[0090] When the upper limit value of the prediction range of the load data to be predicted is less than the reference value for adjusting the prediction range of the load data to be predicted, the upper limit value of the prediction range of the load data to be predicted is set equal to the reference value. When the lower limit value of the prediction range of the load data to be predicted is greater than the reference value for adjusting the prediction range of the load data to be predicted, the lower limit value of the prediction range of the load data to be predicted is set equal to the reference value. This adjustment is to ensure that the prediction range of each load data to be predicted can cover the reference value for adjusting the prediction range of the load data to be predicted.

[0091] For the th load data to be predicted, if the upper limit value of the prediction range of the th load data to be predicted , let the upper limit value of the prediction range of the th load data to be predicted ; if the lower limit value of the prediction range of the th load data to be predicted , let the lower limit value of the prediction range of the th load data to be predicted .

[0092] By adjusting the prediction range of the load data to be predicted through the above reference value, compared with the traditional method of adjusting the prediction range of the load data to be predicted only according to the previous load data, it can better adapt to the complex situations such as noise and fluctuations in the actual construction data, make the adjustment of the prediction range more conform to the real law reflected by the historical data, ensure that the prediction range can be within a relatively accurate and reasonable interval, avoid deviations in the prediction range caused by unreasonable reference values, help to more accurately predict the load data of the drilling machine, timely discover potential abnormal situations, and ensure the smooth progress of the construction.

[0093] S3: Use the adjusted prediction range of the load data to be predicted to predict the load data to be predicted and give an early warning of the fault status of the drilling machine according to the prediction result.

[0094] In one embodiment, the method for using the adjusted prediction range of the load data to be predicted by the HTFE algorithm to predict the load data to be predicted is as follows:

[0095]

[0096] In the formula, is the predicted value of the th load data to be predicted, is the upper limit value of the adjusted prediction range of the th load data to be predicted, is the lower limit value of the adjusted prediction range of the th load data to be predicted, , , are all parameters of the HTFE prediction algorithm, , , are determined according to the historical trend of the th load data to be predicted.

[0097] In one embodiment, , , are determined as follows:

[0098] Obtain the th load data before the load data to be predicted, , , is the total number of historical load data of the th load data to be predicted set, (empirical value);

[0099] Analyze the change trend of these 100 load data:

[0100] If it is monotonically increasing, it means that the historical trend of the th load data to be predicted is a growth type. At this time:

[0101] , , ;

[0102] If it is monotonically decreasing, the historical trend of the th load data to be predicted is a decay type. At this time:

[0103] , , ;

[0104] If it is neither monotonically increasing nor monotonically decreasing, at this time:

[0105] The historical trend of the th load data to be predicted is an irregular type, , , .

[0106] In one embodiment, the method for warning the fault state of the drilling machine according to the prediction result is:

[0107] Judging abnormal load data: Obtain the normalized value of the difference between the predicted value and the actual value of each load data. If the normalized value is greater than the preset difference threshold of 0.96 (empirical value), the load data is abnormal load data. If the normalized value is not greater than the preset difference threshold of 0.96, the load data is normal load data.

[0108] Continuously monitor. If N abnormal load data appear within a certain minute, issue a warning to notify relevant personnel to maintain the drilling machine. N is 5 (empirical value).

[0109] The present invention also provides a construction data processing system based on big data. The construction data processing system includes a memory and a processor. A computer program is stored on the memory, and the processor executes the computer program to implement the steps of any one of the construction data processing methods described above.

[0110] In the description of this specification, the meanings of "a plurality of" and "several" are at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0111] Although this specification has shown and described multiple embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, alterations and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted in the process of practicing the present invention.

Claims

1. A construction data processing method based on big data, characterized in that: include: Collect the load data of the drilling machine during the construction process and the current data of the motor at the same time, take each load data as the load data to be predicted, and use the HTFE prediction algorithm to obtain the prediction range of each load data to be predicted; Set multiple historical load data for each load data to be predicted, calculate the prominence of each historical load data, and evaluate the credibility of each historical load data according to the prominence of each historical load data and the load data in the adjacent data segment of the historical load data, as well as the change trend of the corresponding current data in the adjacent data segment; the credibility of each historical load data satisfies the following relationship: ； In the formula, For the The load data to be predicted The credibility of historical load data, For the The load data to be predicted The prominence of historical load data, For the The load data to be predicted The value of the current data corresponding to the collection time of the historical load data, For the The load data to be predicted The mean of the adjacent data segments of the historical load data, For the The load data to be predicted The value of historical load data, For the The load data to be predicted The average value of the current data in the adjacent data segment of the historical load data, is the absolute value symbol, is the natural exponential function; The calculation formula for the prominence of each historical load data is: ； In the formula, For the The load data to be predicted The prominence of historical load data, For the The load data to be predicted The prediction error of historical load data is For the The load data to be predicted The value of historical load data, For the The load data to be predicted The mean of the adjacent data segments of the historical load data, is a hyperparameter, For the The load data to be predicted The change of historical load data, The change in the historical load data is equal to The historical load data and The absolute value of the difference between the historical load data; Based on the credibility of all historical load data of the load data to be predicted, a reference value for adjusting the prediction range of the load data to be predicted is determined, the prediction range of the load data to be predicted is adjusted using the reference value, the prediction value of the load data to be predicted is calculated according to the adjusted prediction range of the load data to be predicted, and the abnormal load data is determined according to the difference between the predicted value and the actual value of each load data, so as to realize the processing of the data of the drilling machine in the construction process.

2. The construction data processing method based on big data according to claim 1 is characterized in that: The reference value of the prediction range of the load data to be predicted is adjusted to satisfy the following relationship: ； In the formula, To adjust the The reference value of the prediction range of the load data to be predicted, For the The total number of historical load data to be predicted, For the The load data to be predicted The credibility of historical load data, For the The sum of the credibility of all historical load data of the load data to be predicted, For the The load data to be predicted The actual value of historical load data.

3. The construction data processing method based on big data according to claim 1 is characterized in that: The method for obtaining the adjacent data segment of historical load data is as follows: Each historical load data and a plurality of load data immediately before the collection time are combined to form a contiguous data segment of the historical load data.

4. The construction data processing method based on big data according to claim 2 is characterized in that: The method for adjusting the prediction range of the load data to be predicted using the reference value is: Jordi The upper limit of the prediction range of the load data to be predicted is less than , The upper limit of the prediction range of the load data to be predicted is equal to ; Jordi The lower limit of the forecast range of the load data to be forecasted is greater than , The lower limit of the forecast range of the load data to be forecasted is equal to .

5. The construction data processing method based on big data according to claim 4 is characterized in that: The method for calculating the predicted value of the load data to be predicted according to the predicted range after the load data to be predicted is: ； In the formula, For the The predicted value of the load data to be predicted, For the The upper limit value of the forecast range after the load data to be forecasted is adjusted. For the The lower limit of the forecast range after the load data to be forecasted is adjusted. , , These are the parameters of the HTFE prediction algorithm.

6. The construction data processing method based on big data according to claim 5 is characterized in that: The method for determining abnormal load data based on the difference between the predicted value and the actual value of each load data is: The normalized value of the difference between the predicted value and the actual value of each load data is obtained. If the normalized value is greater than a preset difference threshold, the load data is abnormal load data.

7. The construction data processing method based on big data according to claim 6 is characterized in that: After determining the abnormal load data based on the difference between the predicted value and the actual value of each load data, the following operations are performed: Continuous monitoring: if N abnormal load data appear within a minute, an early warning will be issued to notify relevant personnel to perform maintenance on the drilling machine. N is the preset value.

8. A construction data processing system based on big data, characterized in that: The construction data processing system includes a memory and a processor, wherein a computer program is stored in the memory, and the processor executes the computer program to implement the steps of the construction data processing method according to any one of claims 1 to 7.

Images