An all-weather road construction safety early warning device

Abstract

The present invention relates to the technical field of construction environment abnormal pattern recognition, and specifically to an all-weather road construction safety early warning device. The present invention first obtains construction environment data, types of construction accidents and accident levels; further constructs the construction environment data of the same type of construction accident on the same day into an FP tree; further, according to the rule confidence in the FP tree, combined with the deviation characteristics of the construction environment data and the corresponding benchmark value, and combined with the change similarity characteristics of the target class data, obtains the corrected correlation degree between the target class data and the target class accident; further, according to all the corrected correlation degrees of each type of construction environment data, the accident level of the construction accident, the deviation characteristics of each current construction environment data, and the change characteristics of the initial danger index in the current preset historical neighborhood, obtains the current corrected danger index; finally, according to the current corrected danger index, an early warning is issued, and a safety early warning is accurately issued from the perspective of environmental factors to improve construction safety.

Description

Technical Field

[0001] The invention relates to the technical field of abnormal pattern recognition in construction environments, and in particular to an all-weather road construction safety early warning device. Background Art

[0002] The construction safety early warning system can monitor various potential risks at the construction site in real time to prevent accidents. By analyzing equipment failures or various unsafe factors, the system can issue alarms in time to help workers and managers take prompt measures to avoid accidents and injuries.

[0003] However, existing early warning systems or equipment usually only monitor construction equipment and often ignore the impact of environmental factors on construction safety. However, some environmental factors such as extreme weather such as high temperature, severe cold, strong winds, and heavy rain have an important impact on the safety of the construction site. At the same time, due to the complexity and diversity of environmental factors, it is difficult for existing early warning systems to accurately analyze the correlation characteristics between environmental factors and accidents, which further limits its application effect in construction safety management and leads to an unsatisfactory early warning model. Summary of the invention

[0004] In order to solve the technical problem that the existing early warning device does not accurately analyze environmental factors, resulting in an unsatisfactory early warning model, the purpose of the present invention is to provide an all-weather road construction safety early warning device, and the technical solution adopted is as follows:

[0005] An all-weather road construction safety early warning device comprises a device body, the device is used to implement an all-weather road construction safety early warning method, the method comprising:

[0006] Obtain construction environment data, types of construction accidents and accident levels on the day of each construction accident and in the preset historical neighborhood;

[0007] The range of each type of construction environment data on the day of all construction accidents is equally divided into a preset number of levels, and the level of each type of construction environment data is determined; based on the level of the construction environment data on the day of the construction accident, the construction environment data on the day of the same type of construction accidents are constructed into an FP tree; according to the distribution characteristics of each type of construction environment data in the second preset historical neighborhood, the benchmark value of each type of construction environment data is obtained;

[0008] Select any one of the construction environment data as the target class data; select any one of the construction accidents as the target class accident; obtain the degree of association between the target class data and the target class accident according to the rule confidence of the target class data of each level in the FP tree of the target class accident, combined with the deviation characteristics of the construction environment data in each level of the target class data from the corresponding benchmark value; obtain the corrected degree of association between the target class data and the target class accident according to the similar characteristics of the changes of the target class data within the preset historical neighborhood of each construction accident in the target class accident, combined with the corresponding degree of association;

[0009] According to all the modified correlation degrees of each type of the construction environment data, combined with the accident level of the construction accident, the hazard parameter of each type of the construction environment data is obtained; according to the deviation characteristics of each type of the construction environment data from the benchmark value on the current construction day, combined with the hazard parameter, the initial hazard index of the current construction day is obtained; according to the change characteristics of the initial hazard index within the preset historical neighborhood of the current construction day, combined with the initial hazard index of the current construction day, the modified hazard index of the current construction day is obtained;

[0010] An early warning is issued based on the modified risk indicators for the current construction day.

[0011] Furthermore, the method for obtaining the reference value includes:

[0012] The construction environment data is divided into lower limit data and upper limit data; the lower limit data at least includes the minimum temperature; the upper limit data at least includes the maximum temperature, wind speed and precipitation;

[0013] Obtain the construction environment data within a second preset historical neighborhood of the local road construction; obtain a benchmark value for each type of the construction environment data through a preset percentile; wherein the lower limit data of the preset percentage case is not lower than the corresponding benchmark value; the upper limit data of the preset percentage case is not higher than the corresponding benchmark value; the preset percentage case and the preset percentile are the same.

[0014] Furthermore, the method for obtaining the degree of association includes:

[0015] Select any level of the target class data as the target level; when the target class data is the lower limit class data, in the FP tree of the target class accident, the data of the target class data of the target level that is lower than the corresponding reference value is recorded as deviation data; when the target class data is the upper limit class data, in the FP tree of the target class accident, the data of the target class data of the target level that is higher than the corresponding reference value is recorded as deviation data;

[0016] In the FP tree of the target class accident, the sub-correlation degree between the target class data of the target level and the target class accident is obtained according to the rule confidence between the target class data of the target level and the target class accident, combined with the degree of deviation of the deviation data from the corresponding reference value; the rule confidence and the degree of deviation are both positively correlated with the sub-correlation degree;

[0017] The degree of association between the target class data and the target class accident is obtained according to the concentrated characteristics of the sub-association degrees between the target class data of all levels and the target class accident.

[0018] Furthermore, the method for obtaining the modified correlation degree includes:

[0019] The target class data within a preset historical neighborhood of each construction accident in the target class accident is constructed into a time series variation sequence; and the time series variation sequence is arbitrarily combined into a binary group;

[0020] According to the similar features of the time-series change sequences in all the binary groups and in combination with the corresponding correlation degree, the corrected correlation degree between the target type data and the target type accident is obtained; the similar features of the time-series change sequences in the binary groups and the correlation degree are both positively correlated with the corrected correlation degree.

[0021] Furthermore, the method for obtaining the risk parameter includes:

[0022] The average value of the product of the mean value of the accident level of each construction accident and the corrected correlation degree of the target type data is used as the danger parameter of the target type data; the accident level and the accident severity are positively correlated.

[0023] Furthermore, the method for obtaining the initial risk indicator includes:

[0024] Obtain the deviation degree of each construction environment data on the current construction day;

[0025] The sum of the product of the deviation degree of each construction environment data and the hazard parameter on the current construction day is used as the initial hazard index of the current construction day.

[0026] Furthermore, the method for obtaining the modified risk index includes:

[0027] Obtaining the change rate of the initial hazard index in the preset historical neighborhood of the current construction day; obtaining the reference weight of each historical construction day according to the time difference between the historical construction day in the preset historical neighborhood and the current construction day; the time difference is negatively correlated with the reference weight;

[0028] According to the initial risk index of the current construction day, the change rate of the initial risk index of each corresponding historical construction day and the reference weight, the revised risk index of the current construction day is obtained; the initial risk index, the change rate of the initial risk index and the reference weight are all positively correlated with the revised risk index.

[0029] Furthermore, the method for providing early warning according to the modified risk index of the current construction day includes:

[0030] If the modified risk index of the current construction day is greater than the preset warning threshold, a warning alarm is issued.

[0031] Furthermore, the length of the historical neighborhood is preset to be 10 days.

[0032] Furthermore, the preset number is 5.

[0033] The present invention has the following beneficial effects:

[0034] The present invention first obtains the construction environment data, types of construction accidents and accident levels on the day of each construction accident and a preset historical neighborhood, so as to provide a basis for subsequent analysis; further determines the level of each construction environment data, so as to facilitate the construction of an FP tree; further based on the level of the construction environment data on the day of the construction accident, the construction environment data on the day of the same type of construction accident are constructed into an FP tree, so as to facilitate the analysis of the correlation between environmental factors and construction accidents through the FP tree; further obtains the baseline value of each type of construction environment data, so as to provide a baseline for the subsequent analysis of the abnormality of the construction environment data; further combines the FP tree and the abnormality degree of the construction environment data to obtain the correlation degree between the target type data and the target type accident, characterizes the correlation characteristics between the target type data and the target type accident, and provides a basis for subsequent safety The invention provides a basis for early warning; further, from the perspective of the similarity of changes in environmental factors before construction accidents, the degree of correlation is corrected, and the corrected degree of correlation between target class data and target class accidents is obtained to improve the accuracy of the degree of correlation; further, according to all the corrected degrees of correlation of each type of construction environment data, combined with the accident level of the construction accident, the danger parameters of each type of construction environment data are obtained to quantify the danger of each environmental factor; further, the initial danger index of the current construction day is obtained, and the cumulative trend of construction danger is revealed through the change characteristics of the initial danger index in the preset historical neighborhood of the current construction day, and the corrected danger index of the current construction day is obtained, which provides a more reliable basis for the safety early warning device to issue an early warning; finally, an early warning is issued according to the corrected danger index of the current construction day to ensure construction safety. The present invention uses the FP tree and the similarity of changes in environmental data before the construction accident to analyze the correlation between construction environmental factors and construction accidents, evaluates the danger based on the deviation characteristics of the current environmental data, accurately performs safety early warning from the perspective of environmental factors, and improves construction safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 A flow chart of an all-weather road construction safety early warning method provided by an embodiment of the present invention;

[0037] Figure 2 The present invention is a flowchart of a method for obtaining a degree of association provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation, structure, features and effects of an all-weather road construction safety warning device proposed according to the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0039] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0040] The specific scheme of the all-weather road construction safety warning device provided by the present invention is described in detail below with reference to the accompanying drawings.

[0041] The embodiment of the present invention proposes an all-weather road construction safety warning device, which includes a device body, including a data acquisition module, a data analysis module and an early warning module; the data acquisition module, the data analysis module and the early warning module cooperate with each other to jointly implement an all-weather road construction safety early warning method, please refer to Figure 1 , which shows a flow chart of an all-weather road construction safety early warning method provided by an embodiment of the present invention, specifically comprising:

[0042] Step S1: Obtain the construction environment data, types of construction accidents and accident levels on the day of each construction accident and in a preset historical neighborhood.

[0043] In an embodiment of the present invention, considering that environmental factors before a construction accident may have an impact on road construction, and the correlation characteristics between different types of construction accidents and different types of environmental factors may also be different, the construction environment data, types of construction accidents and accident levels on the day of each construction accident and in a preset historical neighborhood are obtained to provide a basis for subsequent analysis.

[0044] In one embodiment of the present invention, construction logs during road construction are collected, semantic recognition is performed on the information in the construction logs and construction accident records, the required data is obtained, the data is cleaned, and stored in the table structure of the database for easy data retrieval. Construction environment data is divided into lower limit data and upper limit data; the lower limit data at least includes the minimum temperature; the upper limit data at least includes the maximum temperature, wind speed and precipitation; the preset length of the historical neighborhood is 10 days.

[0045] It should be noted that the lower limit data represents environmental factors data whose lower values ​​are more unfavorable to road construction safety; the upper limit data represents environmental factors data whose higher values ​​are more unfavorable to road construction safety; in other embodiments of the present invention, implementers may also choose to monitor other types of environmental data, and semantic recognition is already a prior art and will not be elaborated here.

[0046] Step S2: Divide the range of each type of construction environment data on the day of all construction accidents into a preset number of levels, and determine the level of each construction environment data; based on the level of the construction environment data on the day of the construction accident, construct the construction environment data on the day of the same type of construction accidents into an FP tree; obtain the baseline value of each type of construction environment data according to the distribution characteristics of each type of construction environment data in the second preset historical neighborhood.

[0047] Considering that FP tree (Frequent Pattern Tree) is a commonly used data mining method, FP tree can process multidimensional association rules by compressing data into a tree structure, and is very effective for analyzing the relationship between complex and diverse environmental factors and construction accident types, the construction environment data of the same type of construction accidents are constructed into an FP tree; considering that the construction environment data is specific numerical information, directly constructing an FP tree will lead to too many categories and a large FP tree structure, resulting in low calculation efficiency, and the information is too scattered, so the range of each construction environment data on the day of all construction accidents is divided into a preset number of levels, the level of each construction environment data is determined, and similar environment data are divided into one level, which is more conducive to the construction of FP tree, so based on the level of the construction environment data on the day of the construction accident, the construction environment data of the same type of construction accidents on the day of the construction accident are constructed into an FP tree.

[0048] Preferably, in one embodiment of the present invention, by collecting each type of construction environment data on the day of all construction accidents, the range of each type of construction environment data is determined, the range of the data is divided into 5 equal parts, each part is a level, and the construction environment data on the day of the construction accident of the same type of construction accidents are constructed into an FP tree.

[0049] It should be noted that the method of constructing an FP tree is a technical means well known to those skilled in the art and will not be described in detail here.

[0050] In order to analyze the abnormal characteristics of environmental data and determine the impact of abnormal environmental factors on construction accidents during road construction, it is first necessary to determine the baseline value of the construction environment data, so as to determine the normal range of the data. Therefore, according to the distribution characteristics of each type of construction environment data in the second preset historical neighborhood, the baseline value of each type of construction environment data is obtained by analyzing the historical data, which is convenient for subsequent analysis of the correlation characteristics between the abnormal environmental data and construction accidents, and finally early warning.

[0051] Preferably, in one embodiment of the present invention, considering that percentiles can represent the typical distribution of data and exclude the influence of extreme values, thereby improving the reliability of the benchmark value, the construction environment data in the second preset historical neighborhood of the local road construction is obtained; the benchmark value of each type of construction environment data is obtained by the preset percentile;

[0052] The lower limit data of the preset percentage case is not lower than the corresponding benchmark value; the upper limit data of the preset percentage case is not higher than the corresponding benchmark value; the preset percentage case and the preset percentile are the same.

[0053] As an example, the second preset history neighborhood length is 90 days, and the preset percentage case and preset percentile are both 95%. For lower limit data, such as minimum temperature, 95% of the minimum temperatures are not lower than the minimum temperature benchmark value; for upper limit data, such as maximum temperature, 95% of the maximum temperatures are not higher than the maximum temperature benchmark value.

[0054] Step S3: Select any kind of construction environment data as the target class data; select any kind of construction accident as the target class accident; obtain the degree of correlation between the target class data and the target class accident according to the rule confidence of each level of the target class data in the FP tree of the target class accident, combined with the deviation characteristics of the construction environment data in each level of the target class data from the corresponding benchmark value; obtain the corrected degree of correlation between the target class data and the target class accident according to the similar characteristics of the changes in the target class data within the preset historical neighborhood of each construction accident in the target class accident, combined with the corresponding degree of correlation.

[0055] Considering that there are many types of construction accidents and construction environment data during road construction, we first select any type of construction environment data as the target data; and select any type of construction accident as the target accident to facilitate analysis one by one.

[0056] Considering that the rule confidence contained in the FP tree represents the strength of the association rule, that is, the degree of association; at the same time, the more times the construction environment data anomalies and construction accidents occur at the same time, the greater the degree of anomaly of the construction environment data, which means that the correlation between environmental factors and construction accidents is stronger, and the deviation characteristics of the construction environment data relative to the corresponding benchmark values ​​reflect the abnormal situation of the construction environment data. Therefore, according to the rule confidence of each level of target class data in the FP tree of target class accidents, combined with the deviation characteristics of the construction environment data and the corresponding benchmark values ​​in each level of target class data, the degree of association between the target class data and the target class accidents is obtained, and the association characteristics between the target class data and the target class accidents are characterized, providing a basis for subsequent safety warnings.

[0057] Preferably, in one embodiment of the present invention, the method for obtaining the degree of association includes:

[0058] See also Figure 2 , which shows a flow chart of a method for obtaining a degree of association provided by an embodiment of the present invention, specifically comprising:

[0059] Step S301: Select any level of the target class data as the target level; when the target class data is the lower limit class data, in the FP tree of the target class accident, the data in the target class data of the target level that is lower than the corresponding benchmark value is recorded as deviation data; when the target class data is the upper limit class data, in the FP tree of the target class accident, the data in the target class data of the target level that is higher than the corresponding benchmark value is recorded as deviation data.

[0060] Considering that there are multiple levels of target class data, and each level has different association characteristics with target class accidents, any level of target class data is selected as the target level and analyzed one by one;

[0061] Taking into account that different categories of target data deviate from the normal data range in different directions, when the upper limit data is higher than the benchmark value, it is abnormal; and when the lower limit data is lower than the benchmark value, it is abnormal. Therefore, the deviation data is determined according to the type of target data.

[0062] Step S302: In the FP tree of the target class accident, the sub-correlation degree between the target class data of the target level and the target class accident is obtained according to the rule confidence of the target class data of the target level and the target class accident, combined with the degree of deviation of the deviation data from the corresponding benchmark value.

[0063] Among them, rule confidence and deviation degree are both positively correlated with the sub-association degree.

[0064] Considering that the greater the rule confidence, the greater the probability that the target class data of the target level and the target class accident will appear at the same time, and the greater the deviation of the deviation data from the benchmark value, the greater the abnormality of the construction environment data, and the more likely it is to be the cause of the construction accident, so the sub-correlation degree is obtained in this way.

[0065] As an example, the calculation formula of the sub-association degree includes:

[0066] ;

[0067] Among them, x represents the type number of the target accident; y represents the type number of the target data; n represents the level of the data; It represents the degree of sub-correlation between the xth target class accident and the yth target class data at the nth level; Represents the rule confidence between the x-th target class accident and the y-th target class data at the n-th level; The number of deviation data of the yth target class data at the nth level in the FP tree of the xth target class accident; Indicates the sequence number of the deviation data; Represents the benchmark value of the y-th target class data; The data value of the mth deviation data of the yth target class data at the nth level in the FP tree of the xth target class accident; Indicates taking the absolute value; It represents the degree of deviation corresponding to the m-th deviation data of the y-th target class data at the n-th level in the FP tree of the x-th target class accident.

[0068] In the calculation formula of the sub-correlation degree, the absolute deviation characteristics of the deviation data from the benchmark value are expressed by the absolute value of the difference, and then the relative deviation degree of the deviation data is expressed by dividing it by the benchmark value, avoiding the interference of data magnitudes of data in different construction environments, thereby highlighting the deviation characteristics of the data; the deviation degree and rule confidence are integrated by multiplication to obtain the sub-correlation degree.

[0069] It should be noted that when there is no deviation data, The item is 1; the rule confidence can be directly obtained through the FP tree, which will not be repeated here.

[0070] Step S303: Obtain the degree of association between the target class data and the target class accidents according to the concentrated features of the sub-association degrees between the target class data of all levels and the target class accidents.

[0071] After obtaining the sub-correlation degree between the target class data and the target class accidents at each level, it is necessary to integrate all the sub-correlation degrees to obtain the correlation degree; considering that the centralized characteristics of the data represent the overall characteristics of the data set, the correlation degree between the target class data and the target class accidents is obtained based on the centralized characteristics of the sub-correlation degrees between the target class data and the target class accidents at all levels.

[0072] As an example, the concentrated feature of the data is represented by an average value, and the average value of the sub-correlation degrees of the target class data of all levels and the target class accidents is used as the correlation degree between the target class data and the target class accidents.

[0073] In other embodiments of the present invention, implementers may also combine the mean, median and mode to express the concentrated characteristics of the data, such as weighting the mean, median and mode of the sub-correlation degrees with weights of 0.5, 0.2 and 0.3 to sum them up as the correlation degree.

[0074] In the embodiment of the present invention, it is taken into account that the FP tree can only represent the correlation between the environmental factors on the day of the accident and the accident, but in some cases, the occurrence of the accident is not only related to the environmental factors on the day, but may also be the result of being affected by a certain environmental factor for consecutive days; if the change of a certain environmental factor before a certain type of accident occurs is more consistent, it means that the correlation between this type of accident and this type of environmental factor is greater. Therefore, according to the similar characteristics of the changes in the target type data within the preset historical neighborhood of each construction accident in the target type accident, combined with the corresponding correlation degree, the corrected correlation degree between the target type data and the target type accident is obtained, and on the basis of the FP tree, the correlation between environmental factors and accidents in the historical data is integrated to improve the accuracy of the correlation degree, so as to facilitate the final accurate early warning.

[0075] Preferably, in one embodiment of the present invention, considering the similarity of analysis between time series data sequences, it is relatively easy to express the similar characteristics of changes in target class data of each construction accident of the same type, so the target class data within the preset historical neighborhood of each construction accident in the target class accident is constructed as a time series change sequence; the time series change sequences are arbitrarily combined into two-tuples;

[0076] Considering that the more obvious the similar features of all binary groups are, the more likely environmental factors are to be the cause of construction accidents, and the more obvious the correlation features are, the corrected correlation degree between the target data and the target accidents is obtained based on the similar features of the time series change sequences in all binary groups and the corresponding correlation degree; the similar features and correlation degree of the time series change sequences in the binary groups are positively correlated with the corrected correlation degree.

[0077] As an example, suppose that in the xth construction accident, the change sequence of the minimum temperature in the 10 days before the pth accident is , there are s times of the xth type of construction accident, so we can get Considering that the greater the DTW similarity of two time series data, the greater the similarity between the two time series data, the similarity corresponding to each bigram is obtained through the DTW algorithm. ; The calculation formula for the modified correlation degree includes:

[0078] ;

[0079] Among them, x represents the type number of the target accident; y represents the type number of the target data; Indicates the degree of correlation between the x-th target type accident and the y-th target type data; It indicates the degree of correlation between the x-th target type accident and the y-th target type data; Indicates a preset positive parameter; represents the number of tuples of the x-th target type of accident, ; d represents the sequence number of the binary group; Represents the similarity between the time-series change sequences within the d-th bigram.

[0080] In the calculation formula of the modified correlation degree, the similarity corresponding to each bigram is obtained by the DTW algorithm, which indicates the similarity characteristics of the time series change sequence within the bigram. The similarity characteristics of all bigrams are integrated by the average value. In this example, That is, when the average DTW similarity of all binary groups is greater than 0.6, it indicates that the correlation between the x-th construction accident and the lowest temperature is large, and the correlation degree is positively corrected to improve the accuracy of the corrected correlation degree in characterizing the correlation characteristics between environmental factors and construction accidents.

[0081] It should be noted that, in other embodiments of the present invention, the implementer may also obtain the cosine similarity of the two-tuple and convert it into When the average cosine similarity of all binary groups is positive, it means that the greater the similarity of the binary groups, the stronger the correlation between environmental factors and construction accidents, and the correlation degree is amplified and corrected; when the average cosine similarity is negative, the correlation degree is reduced and corrected; at the same time, the correction amplitude parameter can be added, such as , , Represents the cosine similarity between the time series of the d-th binary group; at this time, the restriction for The correction range is 0.5 times to 1.5 times.

[0082] It should be noted that the DTW algorithm and the cosine similarity algorithm are both well-known technical means to those skilled in the art, and will not be described in detail here.

[0083] It should be noted that if a certain type of construction accident has only one record, since it is impossible to compare the relevant data of similar construction accidents, the correlation degree of this type of construction accident will not be corrected, and the correlation degree will be directly used as the corrected correlation degree.

[0084] Step S4: According to all the revised correlation degrees of each type of construction environment data, combined with the accident level of the construction accident, the hazard parameters of each type of construction environment data are obtained; according to the deviation characteristics of each type of construction environment data from the benchmark value on the current construction day, combined with the hazard parameters, the initial hazard index of the current construction day is obtained; according to the change characteristics of the initial hazard index within the preset historical neighborhood of the current construction day, combined with the initial hazard index of the current construction day, the revised hazard index of the current construction day is obtained.

[0085] Taking into account that the hazards of construction accidents caused by different environmental factors are different, the accident level represents the degree of harm of the construction accident, and the corrected correlation degree reflects the degree to which the construction accident is affected by environmental factors. Therefore, according to all the corrected correlation degrees of each construction environment data and combined with the accident level of the construction accident, the danger parameters of each construction environment data are obtained, and the danger of each environmental factor is quantified, which is convenient for accurate early warning based on the current environmental factors.

[0086] Preferably, in one embodiment of the present invention, the accident level and the accident severity are positively correlated; considering that the larger the accident level, the larger the corrected correlation degree corresponding to the target class data, indicating that the danger of the construction accident caused by the target class data is higher, so the average value of the product of the mean value of the accident level of each construction accident and the corrected correlation degree of the target class data is taken as the danger parameter of the target class data;

[0087] As an example, referring to the Regulations on Reporting, Investigation and Handling of Production Safety Accidents, there are four levels of accidents, including general accidents at level 1, major accidents at level 2, serious accidents at level 3, and particularly serious accidents at level 4. When calculating the hazard parameters, the level values ​​of the accident levels are used for calculation.

[0088] Taking into account the deviation characteristics of the construction environment data of the current construction day from the baseline value, it reflects the abnormal degree of the current environmental factors, and the hazard parameters represent the danger of construction accidents caused by environmental factors. Therefore, according to the deviation characteristics of each construction environment data of the current construction day from the baseline value, combined with the hazard parameters, the initial hazard index of the current construction day is obtained to characterize the current construction hazard and provide a basis for the safety warning device to issue an early warning.

[0089] Preferably, in one embodiment of the present invention, considering that the greater the deviation of the current construction environment data, the greater the danger caused by the abnormal construction environment data, that is, the greater the danger parameter, the higher the current construction danger, so the deviation degree of each construction environment data of the current construction day is obtained;

[0090] The sum of the deviation degree of each construction environment data on the current construction day and the product of the hazard parameter is used as the initial hazard index of the current construction day.

[0091] It should be noted that when the construction environment data does not belong to deviation data, that is, when the construction environment data is within the normal range, the deviation degree is 0.

[0092] Taking into account the changing characteristics of the initial hazard indicators in the preset historical neighborhood of the current construction day, the cumulative trend of construction hazards can be revealed. Therefore, according to the changing characteristics of the initial hazard indicators in the preset historical neighborhood of the current construction day, combined with the initial hazard indicators of the current construction day, the revised hazard indicators of the current construction day are obtained, and the initial hazard indicators are further revised to provide a more reliable basis for the safety warning device to issue an early warning.

[0093] Preferably, in one embodiment of the present invention, considering that the more the initial risk index is in an increasing trend, the greater the change rate of the initial risk index is, indicating that the risk of the current construction is gradually accumulating, and the higher the risk of the current construction is; and considering that the closer the historical construction date is to the current construction date, the stronger the reference, so different weights are given to the change rate of the initial risk index according to the time difference, based on this:

[0094] Obtain the change rate of the initial hazard index in the preset historical neighborhood of the current construction day; obtain the reference weight of each historical construction day based on the time difference between the historical construction day in the preset historical neighborhood and the current construction day; the time difference is negatively correlated with the reference weight;

[0095] According to the initial risk index of the current construction day, the change rate of the initial risk index of each corresponding historical construction day and the reference weight, the revised risk index of the current construction day is obtained; the initial risk index, the change rate of the initial risk index and the reference weight are all positively correlated with the revised risk index.

[0096] As an example, the calculation formula for the modified risk index includes:

[0097] ;

[0098] Among them, Q represents the modified hazard index of the current construction day; represents the linear normalization function; Indicates the initial hazard index for the current construction day; Represents the number of historical construction days within the preset historical neighborhood. In this example, ; Indicates the chronological number of the historical construction date; represents the rate of change of the initial hazard index on the jth historical construction day; Represents the reference weight of the jth historical construction day.

[0099] In the calculation formula of the modified hazard index, the time difference is negatively correlated by taking the inverse to obtain the reference weight of each historical construction day; When it is greater than 0, it means that within the preset historical neighborhood of the current construction day, the initial hazard index shows an overall upward trend, and the hazard gradually increases. The initial hazard index of the current construction day is further amplified through historical data, and the sensitivity of the modified hazard index is improved.

[0100] It should be noted that, in one embodiment of the present invention, the rate of change of the initial hazard index of each historical construction day is calculated based on the change of the initial hazard index of each historical construction day relative to the adjacent previous day; in other embodiments of the present invention, the implementer can fit a change curve of the initial hazard index, with the horizontal axis being the historical construction day and the vertical axis being the initial hazard index, and obtain the rate of change of the initial hazard index of each historical construction day through the tangent of the corresponding data point on the change curve.

[0101] Step S5: Issue an early warning based on the revised risk index of the current construction day.

[0102] The modified danger index characterizes the dangerous situation of the current construction day, and finally an early warning is issued based on the modified danger index of the current construction day.

[0103] Preferably, in one embodiment of the present invention, if the modified risk index of the current construction day is greater than a preset warning threshold, a warning alarm is issued.

[0104] As an example, the preset warning threshold is 0.7.

[0105] It should be noted that the embodiment of the present invention mainly analyzes the relationship between environmental factors and construction accidents, and thus issues an alarm. In practical applications, it can work in conjunction with existing monitoring and early warning systems for other construction safety factors, such as an early warning system that monitors the vibration of construction equipment, an early warning system that monitors the concentration of harmful gases, etc., to jointly ensure the safety of road construction.

[0106] In summary, in view of the technical problem that the existing early warning device does not accurately analyze environmental factors, resulting in an unsatisfactory early warning model, the present invention provides an all-weather road construction safety early warning device. The present invention first obtains construction environment data, types of construction accidents and accident levels; further constructs the construction environment data of the same type of construction accident on the day into an FP tree; further, according to the rule confidence in the FP tree, combined with the deviation characteristics of the construction environment data and the corresponding benchmark value, and combined with the change similarity characteristics of the target class data, the corrected correlation degree between the target class data and the target class accident is obtained; further, according to all the corrected correlation degrees of each type of construction environment data, the accident level of the construction accident, the deviation characteristics of each current construction environment data, and the change characteristics of the initial danger index in the preset historical neighborhood of the current construction day, the corrected danger index of the current construction day is obtained; finally, an early warning is issued according to the corrected danger index of the current construction day. The present invention uses the FP tree and the similarity of the changes in the environmental data before the construction accident occurs to analyze the correlation between the construction environment factors and the construction accident, evaluate the danger based on the deviation characteristics of the current environmental data, accurately perform safety early warning from the perspective of environmental factors, and improve construction safety.

[0107] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0108] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. An all-weather road construction safety warning device, comprising a device body, characterized in that: The device is used to implement an all-weather road construction safety early warning method, the method comprising: Obtain construction environment data, types of construction accidents and accident levels on the day of each construction accident and in the preset historical neighborhood; The range of each type of construction environment data on the day of all construction accidents is equally divided into a preset number of levels, and the level of each type of construction environment data is determined; based on the level of the construction environment data on the day of the construction accident, the construction environment data on the day of the same type of construction accidents are constructed into an FP tree; according to the distribution characteristics of each type of construction environment data in the second preset historical neighborhood, the benchmark value of each type of construction environment data is obtained; Select any one of the construction environment data as the target class data; select any one of the construction accidents as the target class accidents and analyze them one by one; obtain the degree of association between the target class data and the target class accident according to the rule confidence of the target class data of each level in the FP tree of the target class accident, combined with the deviation characteristics of the construction environment data in each level of the target class data from the corresponding benchmark value; obtain the corrected degree of association between the target class data and the target class accident according to the similar characteristics of the changes of the target class data within the preset historical neighborhood of each construction accident in the target class accident, combined with the corresponding degree of association; According to all the modified correlation degrees of each type of the construction environment data, combined with the accident level of the construction accident, the hazard parameter of each type of the construction environment data is obtained; according to the deviation characteristics of each type of the construction environment data from the benchmark value on the current construction day, combined with the hazard parameter, the initial hazard index of the current construction day is obtained; according to the change characteristics of the initial hazard index within the preset historical neighborhood of the current construction day, combined with the initial hazard index of the current construction day, the modified hazard index of the current construction day is obtained; An early warning is issued based on the modified risk indicators for the current construction day.

2. The all-weather road construction safety early warning device according to claim 1, characterized in that: The method for obtaining the reference value includes: The construction environment data is divided into lower limit data and upper limit data; the lower limit data at least includes the minimum temperature; the upper limit data at least includes the maximum temperature, wind speed and precipitation; Obtain the construction environment data within a second preset historical neighborhood on the day of road construction; obtain a benchmark value for each type of the construction environment data through a preset percentile; wherein the lower limit data of the preset percentage case is not lower than the corresponding benchmark value; the upper limit data of the preset percentage case is not higher than the corresponding benchmark value; the preset percentage case and the preset percentile are the same.

3. The all-weather road construction safety early warning device according to claim 2, characterized in that: The method for obtaining the degree of association includes: Select any level of the target class data as the target level; when the target class data is the lower limit class data, in the FP tree of the target class accident, the data of the target class data of the target level that is lower than the corresponding reference value is recorded as deviation data; when the target class data is the upper limit class data, in the FP tree of the target class accident, the data of the target class data of the target level that is higher than the corresponding reference value is recorded as deviation data; In the FP tree of the target class accident, the sub-correlation degree between the target class data of the target level and the target class accident is obtained according to the rule confidence between the target class data of the target level and the target class accident, combined with the degree of deviation of the deviation data from the corresponding reference value; the rule confidence and the degree of deviation are both positively correlated with the sub-correlation degree; The degree of association between the target class data and the target class accident is obtained according to the concentrated characteristics of the sub-association degrees between the target class data of all levels and the target class accident.

4. The all-weather road construction safety early warning device according to claim 1, characterized in that: The method for obtaining the modified association degree includes: The target class data within a preset historical neighborhood of each construction accident in the target class accident is constructed into a time series variation sequence; and the time series variation sequence is arbitrarily combined into a binary group; According to the similar features of the time-series change sequences in all the binary groups and in combination with the corresponding correlation degree, the corrected correlation degree between the target type data and the target type accident is obtained; the similar features of the time-series change sequences in the binary groups and the correlation degree are both positively correlated with the corrected correlation degree.

5. The all-weather road construction safety early warning device according to claim 1, characterized in that: The method for obtaining the risk parameter includes: The average value of the product of the mean value of the accident level of each construction accident and the corrected correlation degree of the target type data is used as the danger parameter of the target type data; the accident level and the accident severity are positively correlated.

6. The all-weather road construction safety early warning device according to claim 3, characterized in that: The method for obtaining the initial risk index includes: Obtain the deviation degree of each construction environment data on the current construction day; The sum of the product of the deviation degree of each construction environment data and the hazard parameter on the current construction day is used as the initial hazard index of the current construction day.

7. The all-weather road construction safety early warning device according to claim 1, characterized in that: The method for obtaining the modified risk index includes: Obtaining the change rate of the initial hazard index in the preset historical neighborhood of the current construction day; obtaining the reference weight of each historical construction day according to the time difference between the historical construction day in the preset historical neighborhood and the current construction day; the time difference is negatively correlated with the reference weight; According to the initial risk index of the current construction day, the change rate of the initial risk index of each corresponding historical construction day and the reference weight, the revised risk index of the current construction day is obtained; the initial risk index, the change rate of the initial risk index and the reference weight are all positively correlated with the revised risk index.

8. The all-weather road construction safety early warning device according to claim 1, characterized in that: The method for providing early warning based on the modified risk index of the current construction day includes: If the modified risk index of the current construction day is greater than the preset warning threshold, a warning alarm is issued.

9. The all-weather road construction safety early warning device according to claim 1, characterized in that: The default length of the historical neighborhood is 10 days.

10. The all-weather road construction safety early warning device according to claim 1, characterized in that: The preset number is 5.

Images