Computer room safety early warning system and method based on digital twinning

Abstract

The invention discloses a machine room safety early warning system and method based on digital twinning, relates to the field of safety early warning, and carries out early warning on the safety of a machine room from two aspects of environmental factors and human factors. In the aspect of environmental factors, the final BP neural network prediction model is constructed to predict environmental factor data at a future moment, so that whether the future environmental factor data is abnormal or not is known in advance, and whether early warning needs to be carried out or not is judged; in the human factor aspect, the constructed initial personnel classification mapping equation is optimized, so that a final personnel classification mapping equation capable of accurately classifying the characteristic data of the personnel is obtained; then collecting feature data of corresponding entering personnel, classifying the feature data of the entering personnel by adopting the optimized final personnel classification mapping equation, and according to a classification result, obtaining the feature data of the entering personnel; through a dual early warning mechanism, safety accidents in human and environment aspects of the machine room can be effectively prevented.

Description

Technical Field

[0001] The present invention belongs to the field of security warning, and specifically, particularly relates to a computer room security warning system and method based on digital twins. Background Art

[0002] In addition to environmental factors, human factors also affect the security of the computer room. If there is no control and warning of the people entering, some unauthorized people may enter the computer room, which may lead to the leakage of important data in the computer room and some other unpredictable accidents, causing serious losses to the company or related personnel. Summary of the invention

[0003] In response to the problems in the related technology, the present invention proposes a computer room safety warning system and method based on digital twins to overcome the above-mentioned technical problems existing in the existing related technology.

[0004] In order to solve the above technical problems, the present invention is achieved through the following technical solutions:

[0005] The present invention is a computer room safety early warning method based on digital twins, comprising the following steps:

[0006] S1. Collect characteristic data of several authorized personnel entering the room to be warned, and obtain a characteristic data matrix of authorized personnel; collect various types of environmental factor data inside the room to be warned at multiple time points in real time, and obtain a historical environmental factor data matrix;

[0007] S2, using the historical environmental factor data matrix to construct a final BP neural network prediction model; using the historical environmental factor data matrix and the final BP neural network prediction model to predict environmental factor data at future moments to obtain a future environmental factor data matrix; collecting feature data of multiple unauthorized personnel in the room to be warned according to the authorized personnel feature data matrix and merging it with the authorized personnel feature data matrix to obtain a comprehensive personnel feature data matrix;

[0008] S3, constructing an initial personnel classification mapping equation and optimizing the initial personnel classification mapping equation using the comprehensive personnel feature data matrix to obtain a final personnel classification mapping equation;

[0009] S4, determining whether it is necessary to issue a first warning according to the future environmental factor data matrix;

[0010] S5, when the first warning is not issued, using the final personnel classification mapping equation to classify and map the characteristic data of the incoming personnel to determine whether a second warning needs to be issued;

[0011] In the existing computer room, external personnel can be authorized, and authorized personnel can enter the computer room to perform related operations, while unauthorized personnel are not allowed to enter the computer room to perform operations; based on this, the scheme classifies the relevant characteristics of external personnel when they enter the computer room, so as to determine whether they are authorized personnel, and then perform early warning operations; therefore, firstly, the characteristic data of several authorized personnel entering the computer room to be warned are collected, which provides data support for the subsequent construction of equations for mapping and classifying the characteristic data of external personnel; in addition, in addition to the threat from human factors, the safety of the computer room may also come from external environmental factors; based on this, the scheme collects various types of environmental factor data inside the computer room to be warned at multiple time points, for It provides data support for the subsequent construction of prediction models and prediction of environmental factor data at future moments; by constructing the final BP neural network prediction model, a prediction model is provided for subsequent predictions; by predicting the environmental factor data at future moments, it is possible to pre-determine whether there are abnormalities in the future environmental factor data, so that early warning can be given and measures can be taken to prevent it; in addition, by constructing the final personnel classification mapping equation, a classification tool is provided for the subsequent classification of the characteristic data of personnel entering the computer room; when the external environmental factor data is abnormal, a warning is issued; when unauthorized personnel enter the computer room, a secondary warning is issued; through the dual warning mechanism, human and environmental safety accidents can be effectively prevented in the computer room.

[0012] Preferably, the S1 comprises the following steps:

[0013] S11, construct the room to be warned and the corresponding digital twin model, recorded as the room to be warned model; set the personnel feature type set a = {a1, a2, ..., a i ,...,a a′}, a i represents the set i-th personnel feature type, a′ represents the total number of set personnel feature types; cooperate with the personnel feature type set and use the model of the room to be warned to obtain the feature data of the authorized personnel entering the room to be warned, and obtain the authorized personnel feature data matrix as follows,

[0014]

[0015] in, The characteristic data of the jth type of the i-th authorized personnel in the room to be warned, Indicates the total number of authorized personnel in the room to be warned;

[0016] S12, set multiple factors that can affect the safety of the computer room, and obtain the environmental factor type set b = {b1, b2, ..., bi ,...,b b′}, b i represents the i-th environmental factor type set, and b′ represents the total number of environmental factor types set;

[0017] In conjunction with the environmental factor type set, the model of the room to be warned is used to collect the environmental factor data inside the room to be warned at the same interval in real time to obtain the historical environmental factor data matrix as follows,

[0018]

[0019] in, represents the jth type of environmental factor data collected in real time at the i-th time point, Indicates the total number of environmental factor data collected in real time at the corresponding time point;

[0020] The types of personnel characteristics include whether they have clothing and identification, such as work permits, badges, etc.; whether their behavior is normal, such as swiping a card, entering a password or biometric verification, etc.; whether there are any abnormalities in the items they carry, such as laptops, tool kits or security equipment; however, unauthorized personnel carrying unknown items into the computer room may cause alarm; whether there are any abnormalities in facial expressions and eyes: authorized personnel are usually relaxed and natural when entering the computer room, while unauthorized personnel may appear nervous or uneasy; the types of environmental factors include temperature, humidity, and smoke, etc.

[0021] Preferably, S2 comprises the following steps:

[0022] S21, setting a training data ratio and a test data ratio; dividing the historical environmental factor data matrix according to the training data ratio and the test data ratio to obtain a historical environmental factor training data matrix, a historical environmental factor test data matrix, and a historical environmental factor to-be-predicted data matrix;

[0023] S22, constructing an initial BP neural network prediction model; using the historical environmental factor training data matrix and the historical environmental factor test data matrix to train and test the initial BP neural network prediction model respectively; after the training and testing are completed, the final BP neural network prediction model is obtained;

[0024] S23, inputting the historical environmental factor data matrix to be predicted into the final BP neural network prediction model to predict the environmental factor data at several future time points, and obtaining the future environmental factor data matrix c; as follows,

[0025]

[0026] Among them, cij represents the j-th type of environmental factor data at the i-th future time point predicted, and c′ represents the total number of future time points corresponding to the predicted environmental factor data;

[0027] S24, according to the personnel feature type set a={a1, a2, ..., a i ,...,a a′} Collect the characteristic data of unauthorized personnel in multiple warning rooms and compare them with the characteristic data matrix of authorized personnel Collect and merge to obtain a comprehensive personnel feature data matrix

[0028] By dividing the historical environmental factor data matrix, data support is provided for the subsequent training and testing of the initial BP neural network prediction model; after training and testing, the final BP neural network prediction model has good prediction ability for environmental factor data; therefore, the accuracy of using the final BP neural network prediction model to predict environmental factor data at several future time points is better; by collecting feature data of unauthorized personnel in multiple warning rooms, the collected personnel feature data includes both authorized personnel and unauthorized personnel, making the collected personnel feature data more comprehensive.

[0029] Preferably, the S24 comprises the following steps:

[0030] S241, according to the personnel feature type set a={a1, a2, ..., a i ,...,a a′ Collect the characteristic data of unauthorized personnel in multiple rooms to be warned, and obtain the characteristic data matrix of unauthorized personnel as follows,

[0031]

[0032] in, represents the collected feature data of the jth type of unauthorized personnel in the i-th room to be warned, Indicates the total number of unauthorized personnel in the equipment room to be warned;

[0033] S242: The unauthorized personnel feature data matrix and the authorized personnel characteristics data matrix Merge to obtain comprehensive personnel feature data matrix And the corresponding authorization label dataset represents the authorized label data of the ith person in the comprehensive personnel feature data matrix, When the authorization tag data is 0, it means that the person is an unauthorized person entering the room to be warned; when the authorization tag data is 1, it means that the person is an authorized person entering the room to be warned; as follows,

[0034]

[0035] in, Represents the j-th type of characteristic data of the ith person in the comprehensive personnel characteristic data matrix;

[0036] By constructing the authorized label data corresponding to the comprehensive personnel feature data matrix, the correctness of the optimization process direction is guaranteed in the subsequent process of optimizing the initial personnel classification mapping equation.

[0037] Preferably, the S22 comprises the following steps:

[0038] S221, constructing an initial BP neural network prediction model and setting a training error threshold; inputting the historical environmental factor training data matrix into the initial BP neural network prediction model for training; during the training process, when the training error is less than the training error threshold, stopping the training to obtain a trained BP neural network prediction model; otherwise, continuing the training until the training error is less than the training error threshold;

[0039] S222, set a test accuracy threshold; input the historical environmental factor test data matrix into the trained BP neural network prediction model for testing; after the test is completed, obtain the test accuracy; when the test accuracy is greater than or equal to the test accuracy threshold, use the trained BP neural network prediction model as the final BP neural network prediction model; otherwise, return to S221 to continue training the trained BP neural network prediction model until the test accuracy is greater than or equal to the test accuracy threshold;

[0040] Preferably, S3 comprises the following steps:

[0041] S31, according to the comprehensive personnel characteristic data matrix Construct the initial personnel classification mapping equation; as follows,

[0042]

[0043] In the formula, is the dependent variable of the initial personnel classification mapping equation, representing the authorized label data; is the i-th independent variable of the initial personnel classification mapping equation, representing the i-th type of characteristic data; express β1 and β2 represent the first bias and the second bias of the initial personnel classification mapping equation respectively; ceil represents the rounding function;

[0044] S32, the comprehensive personnel feature data matrix Each row of data in is input into the initial personnel classification mapping equation for classification mapping to obtain the authorized label mapping data set d i Indicates that the comprehensive personnel feature data matrix The data in the i-th row is input into the initial personnel classification mapping equation for classification mapping to obtain the authorized label mapping data;

[0045] S33, when there is authorization tag mapping data in the authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set, optimizing the initial personnel classification mapping equation until there is no authorization tag mapping data in the authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set, and obtaining a final personnel classification mapping equation; otherwise, there is no need to optimize the initial personnel classification mapping equation, and the initial personnel classification mapping equation is used as the final personnel classification mapping equation;

[0046] Since there are only two types of classification results in this scheme, the initial personnel classification mapping equation constructed uses an integer function. When the internal function value is less than 1, the entire function value is 0; when the internal function value is greater than or equal to 1 and not more than 2, the entire function value is 1; therefore, the classification of personnel characteristic data can be achieved; then, by performing pre-classification, it is determined whether the relevant parameter values ​​of the constructed initial personnel classification mapping equation are correct; when the mapping accuracy does not meet the requirements, the relevant parameter values ​​of the initial personnel classification mapping equation are optimized.

[0047] Preferably, optimizing the initial personnel classification mapping equation in S33 includes the following steps:

[0048] S331, Construct equations to optimize squirrel populations d i ' represents the i-th squirrel in the squirrel population optimized by the equation, Indicates that the equation optimizes the size of the squirrel population; Sets the maximum number of iterations of the equation to optimize the squirrel population The current number of iterations is are respectively recorded as the first maximum number of iterations and the first current number of iterations; the dimension of the search space for optimizing the squirrel population by the equation is a′+2;

[0049] S332, setting the value intervals of each independent variable coefficient, the first offset and the second offset in the initial personnel classification mapping equation, and obtaining the coefficient value interval set e1, the first offset value value interval [e 21 ,e 22 ] and the second bias value interval [e 31 ,e 32 ]; e 21 、e 22 respectively represent the lower limit and the upper limit of the first bias value; e 31 、e 32 They represent the lower limit and the upper limit of the second bias value respectively; e1 is as follows,

[0050]

[0051] in, Respectively represent the lower limit and upper limit of the value of the i-th independent variable coefficient in the initial personnel classification mapping equation;

[0052] According to the coefficient value interval set, the first bias value interval and the second bias value interval, an equation is set to optimize the initial position of each squirrel in the squirrel population to obtain a first initial position matrix e′; as follows,

[0053]

[0054] Among them, e′ ij represents the component of the initial position of the i-th squirrel in the squirrel population optimized by the equation in the j-th search space dimension; e′ ij , e′ i(a′+1) , e′ i(a′+2) The calculation formulas are as follows:

[0055] e′ i(a′+1) =e 21 +rand i(a′+1) ×(e 22 -e 21 );

[0056] e′ i(a′+2) =e 31 +rand i(a′+2) ×(e 32 -e 31 );

[0057] S333: Mapping a data set according to the authorization tag And the authorized label dataset Constructing equations to optimize the fitness function of the squirrel population as follows

[0058]

[0059] In the formula, β3 is a positive number, indicating the correction parameter;

[0060] S334, start iteration, before iteration, set the first current iteration number Set to 1; in the first round of iterations, the equation is used to optimize the fitness function of the squirrel population Calculate the fitness value of the initial position of each squirrel in the first initial position matrix e′ to obtain a first fitness value set; use the maximum fitness value in the first fitness value set and the corresponding initial position of the squirrel as the first global optimal fitness and the first global optimal position respectively; update the initial position of each squirrel in the first initial position matrix e′ according to the first global optimal fitness and the first global optimal position; after the update is completed, update the first current iteration number Add 1 and enter the next iteration;

[0061] In each other round of iteration, the equation is used to optimize the fitness function of the squirrel population. Calculate the fitness value of each squirrel's position updated in the previous round of iterations to obtain a second fitness value set; use the maximum fitness value in the second fitness value set and the corresponding initial position of the squirrel as the second global optimal fitness and the second global optimal position respectively; update the position of each squirrel updated in the previous round of iterations according to the second global optimal fitness and the second global optimal position; after the update is completed, update the first current iteration number Add 1 and enter the next iteration;

[0062] S335, when When , stop the iteration and get the first final global optimal position; otherwise, continue to iterate until Substitute the first final global optimal position into the initial personnel classification mapping equation to obtain the optimized personnel classification mapping equation; Substitute the comprehensive personnel feature data matrix in S32 Each row of data in is respectively input into the optimized personnel classification mapping equation for classification mapping to obtain an optimized authorization label mapping data set;

[0063] When there is no authorization tag mapping data in the optimized authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set, the optimized personnel classification mapping equation is used as the final personnel classification mapping equation; otherwise, return to S334 to continue iterating until there is no authorization tag mapping data in the optimized authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set;

[0064] The squirrel optimization algorithm simulates the dynamic foraging behavior of squirrels sliding between different tree species to find food. This behavior helps the algorithm to conduct both global exploration and local development during the search process. By simulating the movement of squirrels between different tree species, the algorithm can maintain diversity in the search process and avoid premature convergence to the local optimal solution. Based on the above, the squirrel optimization algorithm is used in this scheme to perform multiple iterative optimizations on multiple independent variable coefficients and biases of the initial personnel classification mapping equation, and the mapping accuracy of the personnel classification mapping equation is used as the fitness function. Therefore, as the iteration proceeds, the mapping accuracy of the personnel classification mapping equation becomes higher and higher until the requirements are met.

[0065] Preferably, S4 comprises the following steps:

[0066] S41, setting the environmental factor data security threshold corresponding to each environmental factor type in the environmental factor type set to obtain an environmental factor security threshold set;

[0067] S42, when the environmental factor data at a certain future moment in the future environmental factor data matrix is ​​greater than or equal to the environmental factor safety threshold corresponding to the environmental factor safety threshold set, a first warning is issued; otherwise, the first warning does not need to be issued;

[0068] The environmental factor data of the future moment predicted previously is compared with the pre-set data threshold. When the environmental factor data exceeds the threshold, it indicates that there is an abnormality in the environmental factor data and an early warning is required.

[0069] Preferably, S5 comprises the following steps:

[0070] S51. When the first warning is not issued, the person who is entering the room to be warned is set as the person to be entered; the feature data of the person to be entered is collected according to the personnel feature type set and in cooperation with the model of the room to be warned, so as to obtain a feature data set of the person to be entered; the feature data set of the person to be entered is substituted into the final personnel classification mapping equation for classification mapping, so as to obtain the authorization label data of the person to be entered;

[0071] S52: When the data of the authorization tag of the person to enter is 0, a second warning is issued; otherwise, no second warning is issued;

[0072] The characteristic data of the personnel entering the room to be warned are classified by the final personnel classification mapping equation optimized previously. When the classification result shows that the personnel are authorized personnel, no warning is required. Otherwise, a warning is required.

[0073] A computer room safety early warning system based on digital twins, comprising an authorized personnel feature collection module, an environmental factor data collection module, a prediction model construction module, a prediction module, a comprehensive personnel feature data construction module, a classification mapping equation construction optimization module, a computer room first early warning determination module, and a computer room second early warning determination module;

[0074] The authorized personnel feature collection module is used to collect feature data of several authorized personnel entering the computer room to be warned, and obtain a matrix of authorized personnel feature data;

[0075] The environmental factor data collection module is used to collect multiple types of environmental factor data inside the warning room at multiple time points in real time to obtain a historical environmental factor data matrix;

[0076] The prediction model building module is used to construct a final BP neural network prediction model using the historical environmental factor data matrix;

[0077] The prediction module is used to cooperate with the historical environmental factor data matrix and use the final BP neural network prediction model to predict the environmental factor data at future moments to obtain the future environmental factor data matrix;

[0078] The comprehensive personnel feature data construction module is used to collect feature data of multiple unauthorized personnel in the warning room according to the authorized personnel feature data matrix and merge it with the authorized personnel feature data matrix to obtain a comprehensive personnel feature data matrix;

[0079] The classification mapping equation construction optimization module is used to construct an initial personnel classification mapping equation and optimize the initial personnel classification mapping equation using the comprehensive personnel feature data matrix to obtain a final personnel classification mapping equation;

[0080] The first warning determination module of the computer room is used to determine whether it is necessary to issue a first warning according to the future environmental factor data matrix;

[0081] The second warning determination module of the computer room is used for classifying and mapping the characteristic data of the personnel to be entered by using the final personnel classification mapping equation to determine whether it is necessary to issue a second warning when the first warning is not issued.

[0082] The present invention has the following beneficial effects:

[0083] 1. The present invention provides early warning for the safety of the computer room from two aspects, environmental factors and human factors. In terms of environmental factors, the final BP neural network prediction model is constructed to predict the environmental factor data at future moments, so as to know in advance whether there are abnormalities in the future environmental factor data and determine whether an early warning is needed; in terms of human factors, the initial personnel classification mapping equation constructed is optimized by pre-collecting the characteristic data of authorized personnel and unauthorized personnel in the computer room, so as to obtain the final personnel classification mapping equation that can accurately classify the characteristic data of personnel; then the characteristic data of the corresponding entering personnel is collected, and the characteristic data of the entering personnel is classified using the optimized final personnel classification mapping equation, and the characteristic data of the entering personnel is entered according to the classification result; through the dual early warning mechanism, the human and environmental safety accidents in the computer room can be effectively prevented.

[0084] 2. The present invention divides the historical environmental factor data matrix to provide data support for subsequent training and testing of the initial BP neural network prediction model; after training and testing, the final BP neural network prediction model obtained has good prediction ability for environmental factor data.

[0085] 3. In the present invention, the squirrel optimization algorithm is used to perform multiple iterative optimizations on multiple independent variable coefficients and biases of the initial personnel classification mapping equation, and the mapping accuracy of the personnel classification mapping equation is used as the fitness function; therefore, as the iteration proceeds, the mapping accuracy of the personnel classification mapping equation becomes higher and higher until the requirements are met.

[0086] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0088] Figure 1 It is a flow chart of a computer room safety early warning method based on digital twins of the present invention. DETAILED DESCRIPTION

[0089] The following will be combined with the drawings in the embodiments of the invention to clearly and completely describe the technical solutions in the embodiments of the invention. Obviously, the described embodiments are only part of the embodiments of the invention, not all of the embodiments. Based on the embodiments in the invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the invention.

[0090] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0091] Embodiment 1

[0092] This embodiment is a computer room safety early warning method based on digital twins, comprising the following steps:

[0093] S1. Collect characteristic data of several authorized personnel entering the room to be warned, and obtain a characteristic data matrix of authorized personnel; collect various types of environmental factor data inside the room to be warned at multiple time points in real time, and obtain a historical environmental factor data matrix;

[0094] The S1 comprises the following steps:

[0095] S11, construct the room to be warned and the corresponding digital twin model, recorded as the room to be warned model; set the personnel feature type set a = {a1, a2, ..., a i ,...,a a′}, a i represents the set i-th personnel feature type, a′ represents the total number of set personnel feature types; cooperate with the personnel feature type set and use the model of the room to be warned to obtain the feature data of the authorized personnel entering the room to be warned, and obtain the authorized personnel feature data matrix as follows,

[0096]

[0097] in, The characteristic data of the jth type of the i-th authorized personnel in the room to be warned, Indicates the total number of authorized personnel in the room to be warned;

[0098] S12, set multiple factors that can affect the safety of the computer room, and obtain the environmental factor type set b = {b1, b2, ..., b i ,...,b b′}, b i represents the i-th environmental factor type set, and b′ represents the total number of environmental factor types set;

[0099] In conjunction with the environmental factor type set, the model of the room to be warned is used to collect the environmental factor data inside the room to be warned at the same interval in real time to obtain the historical environmental factor data matrix as follows,

[0100]

[0101] in, represents the jth type of environmental factor data collected in real time at the i-th time point, Indicates the total number of environmental factor data collected in real time at the corresponding time point;

[0102] S2, using the historical environmental factor data matrix to construct a final BP neural network prediction model; using the historical environmental factor data matrix and the final BP neural network prediction model to predict environmental factor data at future moments to obtain a future environmental factor data matrix; collecting feature data of multiple unauthorized personnel in the room to be warned according to the authorized personnel feature data matrix and merging it with the authorized personnel feature data matrix to obtain a comprehensive personnel feature data matrix;

[0103] The S2 comprises the following steps:

[0104] S21, setting a training data ratio and a test data ratio; dividing the historical environmental factor data matrix according to the training data ratio and the test data ratio to obtain a historical environmental factor training data matrix, a historical environmental factor test data matrix, and a historical environmental factor to-be-predicted data matrix;

[0105] S22, constructing an initial BP neural network prediction model; using the historical environmental factor training data matrix and the historical environmental factor test data matrix to train and test the initial BP neural network prediction model respectively; after the training and testing are completed, the final BP neural network prediction model is obtained;

[0106] The S22 comprises the following steps:

[0107] S221, constructing an initial BP neural network prediction model and setting a training error threshold; inputting the historical environmental factor training data matrix into the initial BP neural network prediction model for training; during the training process, when the training error is less than the training error threshold, stopping the training to obtain a trained BP neural network prediction model; otherwise, continuing the training until the training error is less than the training error threshold;

[0108] S222, set a test accuracy threshold; input the historical environmental factor test data matrix into the trained BP neural network prediction model for testing; after the test is completed, obtain the test accuracy; when the test accuracy is greater than or equal to the test accuracy threshold, use the trained BP neural network prediction model as the final BP neural network prediction model; otherwise, return to S221 to continue training the trained BP neural network prediction model until the test accuracy is greater than or equal to the test accuracy threshold;

[0109] S23, inputting the historical environmental factor data matrix to be predicted into the final BP neural network prediction model to predict the environmental factor data at several future time points, and obtaining the future environmental factor data matrix c; as follows,

[0110]

[0111] Among them, c ij represents the j-th type of environmental factor data at the i-th future time point predicted, and c′ represents the total number of future time points corresponding to the predicted environmental factor data;

[0112] S24, according to the personnel feature type set a={a1, a2, ..., a i ,...,a a′} Collect the characteristic data of unauthorized personnel in multiple warning rooms and compare them with the characteristic data matrix of authorized personnel Collect and merge to obtain a comprehensive personnel feature data matrix

[0113] The S24 comprises the following steps:

[0114] S241, according to the personnel feature type set a={a1, a2, ..., a i ,...,a a′ Collect the characteristic data of unauthorized personnel in multiple rooms to be warned, and obtain the characteristic data matrix of unauthorized personnel as follows,

[0115]

[0116] in, represents the collected feature data of the jth type of unauthorized personnel in the i-th room to be warned, Indicates the total number of unauthorized personnel in the equipment room to be warned;

[0117] S242: The unauthorized personnel feature data matrix and the authorized personnel characteristics data matrix Merge to obtain comprehensive personnel feature data matrix And the corresponding authorization label dataset represents the authorized label data of the ith person in the comprehensive personnel feature data matrix, When the authorization tag data is 0, it means that the person is an unauthorized person entering the room to be warned; when the authorization tag data is 1, it means that the person is an authorized person entering the room to be warned; as follows,

[0118]

[0119] in, Represents the j-th type of characteristic data of the ith person in the comprehensive personnel characteristic data matrix;

[0120] S3, constructing an initial personnel classification mapping equation and optimizing the initial personnel classification mapping equation using the comprehensive personnel feature data matrix to obtain a final personnel classification mapping equation;

[0121] The S3 comprises the following steps:

[0122] S31, according to the comprehensive personnel characteristic data matrix Construct the initial personnel classification mapping equation; as follows,

[0123]

[0124] In the formula, is the dependent variable of the initial personnel classification mapping equation, representing the authorized label data; is the i-th independent variable of the initial personnel classification mapping equation, representing the i-th type of characteristic data; express β1 and β2 represent the first bias and the second bias of the initial personnel classification mapping equation respectively; ceil represents the rounding function;

[0125] S32, the comprehensive personnel feature data matrix Each row of data in is input into the initial personnel classification mapping equation for classification mapping to obtain the authorized label mapping data set d i Indicates that the comprehensive personnel feature data matrix The data in the i-th row is input into the initial personnel classification mapping equation for classification mapping to obtain the authorized label mapping data;

[0126] S33, when there is authorization tag mapping data in the authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set, optimizing the initial personnel classification mapping equation until there is no authorization tag mapping data in the authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set, and obtaining a final personnel classification mapping equation; otherwise, there is no need to optimize the initial personnel classification mapping equation, and the initial personnel classification mapping equation is used as the final personnel classification mapping equation;

[0127] The optimization of the initial personnel classification mapping equation in S33 includes the following steps:

[0128] S331, Construct equations to optimize squirrel populations d i ' represents the i-th squirrel in the squirrel population optimized by the equation, Indicates that the equation optimizes the size of the squirrel population; Sets the maximum number of iterations of the equation to optimize the squirrel population The current number of iterations is are respectively recorded as the first maximum number of iterations and the first current number of iterations; the dimension of the search space for optimizing the squirrel population by the equation is a′+2;

[0129] S332, setting the value intervals of each independent variable coefficient, the first offset and the second offset in the initial personnel classification mapping equation, and obtaining the coefficient value interval set e1, the first offset value value interval [e 21 ,e 22 ] and the second bias value interval [e 31 ,e 32 ]; e 21 、e 22 respectively represent the lower limit and the upper limit of the first bias value; e 31 、e 32 They represent the lower limit and the upper limit of the second bias value respectively; e1 is as follows,

[0130]

[0131] in, Respectively represent the lower limit and upper limit of the value of the i-th independent variable coefficient in the initial personnel classification mapping equation;

[0132] According to the coefficient value interval set, the first bias value interval and the second bias value interval, an equation is set to optimize the initial position of each squirrel in the squirrel population to obtain a first initial position matrix e′; as follows,

[0133]

[0134] Among them, e′ ij represents the component of the initial position of the i-th squirrel in the squirrel population optimized by the equation in the j-th search space dimension; e′ ij , e′ i(a′+1) , e′ i(a′+2) The calculation formulas are as follows:

[0135] e′ i(a′+1) =e 21 +rand i(a′+1) ×(e 22 -e 21 );

[0136] e′ i(a′+2) =e 31 +rand i(a′+2) ×(e 32 -e 31 );

[0137] S333: Mapping a data set according to the authorization tag And the authorized label dataset Constructing equations to optimize the fitness function of the squirrel population as follows

[0138]

[0139] In the formula, β3 is a positive number, indicating the correction parameter;

[0140] S334, start iteration, before iteration, set the first current iteration number Set to 1; in the first round of iterations, the equation is used to optimize the fitness function of the squirrel population Calculate the fitness value of the initial position of each squirrel in the first initial position matrix e′ to obtain a first fitness value set; use the maximum fitness value in the first fitness value set and the corresponding initial position of the squirrel as the first global optimal fitness and the first global optimal position respectively; update the initial position of each squirrel in the first initial position matrix e′ according to the first global optimal fitness and the first global optimal position; after the update is completed, update the first current iteration number Add 1 and enter the next iteration;

[0141] In each other round of iteration, the equation is used to optimize the fitness function of the squirrel population. Calculate the fitness value of each squirrel's position updated in the previous round of iterations to obtain a second fitness value set; use the maximum fitness value in the second fitness value set and the corresponding initial position of the squirrel as the second global optimal fitness and the second global optimal position respectively; update the position of each squirrel updated in the previous round of iterations according to the second global optimal fitness and the second global optimal position; after the update is completed, update the first current iteration number Add 1 and enter the next iteration;

[0142] S335, when When , stop the iteration and get the first final global optimal position; otherwise, continue to iterate until Substitute the first final global optimal position into the initial personnel classification mapping equation to obtain the optimized personnel classification mapping equation; Substitute the comprehensive personnel feature data matrix in S32 Each row of data in is input into the optimized personnel classification mapping equation for classification mapping to obtain an optimized authorization label mapping data set;

[0143] When there is no authorization tag mapping data in the optimized authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set, the optimized personnel classification mapping equation is used as the final personnel classification mapping equation; otherwise, return to S334 to continue iterating until there is no authorization tag mapping data in the optimized authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set;

[0144] S4, determining whether it is necessary to issue a first warning according to the future environmental factor data matrix;

[0145] The S4 comprises the following steps:

[0146] S41, setting the environmental factor data security threshold corresponding to each environmental factor type in the environmental factor type set to obtain an environmental factor security threshold set;

[0147] S42, when the environmental factor data at a certain future moment in the future environmental factor data matrix is ​​greater than or equal to the environmental factor safety threshold corresponding to the environmental factor safety threshold set, a first warning is issued; otherwise, the first warning does not need to be issued;

[0148] S5, when the first warning is not issued, using the final personnel classification mapping equation to classify and map the characteristic data of the incoming personnel to determine whether a second warning needs to be issued;

[0149] The S5 comprises the following steps:

[0150] S51. When the first warning is not issued, the person who is entering the room to be warned is set as the person to be entered; the feature data of the person to be entered is collected according to the personnel feature type set and in cooperation with the model of the room to be warned, so as to obtain a feature data set of the person to be entered; the feature data set of the person to be entered is substituted into the final personnel classification mapping equation for classification mapping, so as to obtain the authorization label data of the person to be entered;

[0151] S52: When the data of the authorization tag of the person to enter is 0, a second warning is issued; otherwise, there is no need to issue a second warning.

[0152] Embodiment 2

[0153] This embodiment discloses a computer room safety warning system based on digital twins, which can implement the method of the above embodiment, including an authorized personnel feature collection module, an environmental factor data collection module, a prediction model construction module, a prediction module, a comprehensive personnel feature data construction module, a classification mapping equation construction optimization module, a computer room first warning determination module, and a computer room second warning determination module;

[0154] The authorized personnel feature collection module is used to collect feature data of several authorized personnel entering the computer room to be warned, and obtain a matrix of authorized personnel feature data;

[0155] The environmental factor data collection module is used to collect multiple types of environmental factor data inside the warning room at multiple time points in real time to obtain a historical environmental factor data matrix;

[0156] The prediction model building module is used to construct a final BP neural network prediction model using the historical environmental factor data matrix;

[0157] The prediction module is used to cooperate with the historical environmental factor data matrix and use the final BP neural network prediction model to predict the environmental factor data at future moments to obtain the future environmental factor data matrix;

[0158] The comprehensive personnel feature data construction module is used to collect feature data of multiple unauthorized personnel in the warning room according to the authorized personnel feature data matrix and merge it with the authorized personnel feature data matrix to obtain a comprehensive personnel feature data matrix;

[0159] The classification mapping equation construction optimization module is used to construct an initial personnel classification mapping equation and optimize the initial personnel classification mapping equation using the comprehensive personnel feature data matrix to obtain a final personnel classification mapping equation;

[0160] The first warning determination module of the computer room is used to determine whether it is necessary to issue a first warning according to the future environmental factor data matrix;

[0161] The second warning determination module of the computer room is used for classifying and mapping the characteristic data of the personnel to be entered by using the final personnel classification mapping equation to determine whether it is necessary to issue a second warning when the first warning is not issued.

[0162] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0163] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. A computer room safety early warning method based on digital twins, characterized in that: The following steps are involved: S1. Collect the characteristic data of several personnel authorized to enter the room to be warned, and obtain the characteristic data matrix of the authorized personnel; Collect various types of environmental factor data inside the warning room at multiple time points in real time to obtain a historical environmental factor data matrix; S2, using the historical environmental factor data matrix to construct a final BP neural network prediction model; using the historical environmental factor data matrix and the final BP neural network prediction model to predict environmental factor data at future moments to obtain a future environmental factor data matrix; collecting feature data of multiple unauthorized personnel in the room to be warned according to the authorized personnel feature data matrix and merging it with the authorized personnel feature data matrix to obtain a comprehensive personnel feature data matrix; S3, constructing an initial personnel classification mapping equation and optimizing the initial personnel classification mapping equation using the comprehensive personnel feature data matrix to obtain a final personnel classification mapping equation; S4, determining whether it is necessary to issue a first warning according to the future environmental factor data matrix; S5, when the first warning is not issued, the final personnel classification mapping equation is used to classify and map the characteristic data of the personnel to be entered to determine whether it is necessary to issue a second warning.

2. According to the digital twin-based computer room safety early warning method of claim 1, it is characterized in that: The S1 comprises the following steps: S11, constructing a computer room to be warned and a corresponding digital twin model, recorded as a computer room to be warned model; setting a personnel feature type set; using the personnel feature type set and the computer room to be warned model to obtain feature data of authorized personnel entering the computer room to be warned, and obtaining a feature data matrix of authorized personnel; S12. Set multiple factors that may affect the safety of the computer room to obtain an environmental factor type set; use the computer room model to collect environmental factor data inside the computer room to be warned at the same interval in real time according to the environmental factor type set to obtain a historical environmental factor data matrix.

3. According to the digital twin-based computer room safety early warning method of claim 2, it is characterized in that: The S2 comprises the following steps: S21, dividing the historical environmental factor data matrix to obtain a historical environmental factor training data matrix, a historical environmental factor test data matrix, and a historical environmental factor to-be-predicted data matrix; S22, constructing an initial BP neural network prediction model; using the historical environmental factor training data matrix and the historical environmental factor test data matrix to train and test the initial BP neural network prediction model respectively; after the training and testing are completed, the final BP neural network prediction model is obtained; S23, inputting the historical environmental factor to-be-predicted data matrix into the final BP neural network prediction model to predict the environmental factor data at several future time points to obtain a future environmental factor data matrix; S24. Collect feature data of multiple unauthorized personnel in the warning room according to the personnel feature type set and merge them with the feature data matrix of authorized personnel to obtain a comprehensive personnel feature data matrix.

4. According to the digital twin-based computer room safety early warning method of claim 3, it is characterized in that: The S22 comprises the following steps: S221, constructing an initial BP neural network prediction model and setting a training error threshold; inputting the historical environmental factor training data matrix into the initial BP neural network prediction model for training; during the training process, when the training error is less than the training error threshold, stopping the training to obtain a trained BP neural network prediction model; otherwise, continuing the training until the training error is less than the training error threshold; S222, set a test accuracy threshold; input the historical environmental factor test data matrix into the trained BP neural network prediction model for testing; after the test is completed, obtain the test accuracy; when the test accuracy is greater than or equal to the test accuracy threshold, use the trained BP neural network prediction model as the final BP neural network prediction model; otherwise, return to S221 to continue training the trained BP neural network prediction model until the test accuracy is greater than or equal to the test accuracy threshold.

5. According to the digital twin-based computer room safety early warning method of claim 4, it is characterized in that: The S24 comprises the following steps: S241, collecting feature data of multiple unauthorized personnel in the computer room to be warned according to the personnel feature type set, and obtaining an unauthorized personnel feature data matrix; S242. Merge the unauthorized personnel feature data matrix and the authorized personnel feature data matrix to obtain a comprehensive personnel feature data matrix and a corresponding authorization label data set; when the authorization label data is 0, it indicates that the person is an unauthorized person entering the room to be warned; when the authorization label data is 1, it indicates that the person is an authorized person entering the room to be warned.

6. According to claim 5, a computer room safety early warning method based on digital twins is characterized in that: The S3 comprises the following steps: S31, constructing an initial personnel classification mapping equation according to the comprehensive personnel feature data matrix; S32, inputting each row of data in the comprehensive personnel feature data matrix into the initial personnel classification mapping equation for classification mapping, and obtaining an authorized label mapping data set; S33. When there exists in the authorization tag mapping data set authorization tag mapping data which is different from the corresponding authorization tag data in the authorization tag data set, the initial personnel classification mapping equation is optimized until there exists in the authorization tag mapping data set authorization tag mapping data which is different from the corresponding authorization tag data in the authorization tag data set, and a final personnel classification mapping equation is obtained; otherwise, there is no need to optimize the initial personnel classification mapping equation, and the initial personnel classification mapping equation is used as the final personnel classification mapping equation.

7. The computer room safety early warning method based on digital twin according to claim 6 is characterized in that: The optimization of the initial personnel classification mapping equation in S33 includes the following steps: S331, constructing an equation to optimize the squirrel population; setting the maximum number of iterations of the equation to optimize the squirrel population to The current number of iterations is They are respectively recorded as the first maximum number of iterations and the first current number of iterations; S332, setting the value intervals of each independent variable coefficient, the first bias value, and the second bias value in the initial personnel classification mapping equation to obtain a coefficient value interval set, a first bias value interval, and a second bias value interval; setting an equation to optimize the initial position of each squirrel in the squirrel population according to the coefficient value interval set, the first bias value interval, and the second bias value interval to obtain a first initial position matrix; S333, constructing an equation to optimize the fitness function of the squirrel population according to the authorized tag mapping data set and the authorized tag data set; S334, start iteration; in each round of iteration, use the equation to optimize the fitness function of the squirrel population to calculate the fitness value of each squirrel position updated in the previous round of iteration and update the position of each squirrel updated in the previous round of iteration again; after the update is completed, the first current iteration number is Add 1 and enter the next iteration; S335, when When , stop the iteration and get the first final global optimal position; otherwise, continue to iterate until until the time; substitute the first final global optimal position into the initial personnel classification mapping equation to obtain the optimized personnel classification mapping equation; input each row of data in the comprehensive personnel feature data matrix in S32 into the optimized personnel classification mapping equation for classification mapping to obtain the optimized authorization label mapping data set; When there is no authorization tag mapping data in the optimized authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set, the optimized personnel classification mapping equation is used as the final personnel classification mapping equation; otherwise, return to S334 and continue to iterate until there is no authorization tag mapping data in the optimized authorization tag mapping data set that is different from the corresponding authorization tag data in the authorization tag data set.

8. The computer room safety early warning method based on digital twin according to claim 7 is characterized in that: The S4 comprises the following steps: S41, setting the environmental factor data security threshold corresponding to each environmental factor type in the environmental factor type set to obtain an environmental factor security threshold set; S42. When the environmental factor data at a certain future moment in the future environmental factor data matrix is ​​greater than or equal to the environmental factor safety threshold corresponding to the environmental factor safety threshold set, a first warning is issued; otherwise, there is no need to issue a first warning.

9. The computer room safety early warning method based on digital twin according to claim 8 is characterized in that: The S5 comprises the following steps: S51. When the first warning is not issued, the person who is entering the room to be warned is set as the person to be entered; the feature data of the person to be entered is collected according to the personnel feature type set and in cooperation with the model of the room to be warned, so as to obtain a feature data set of the person to be entered; the feature data set of the person to be entered is substituted into the final personnel classification mapping equation for classification mapping, so as to obtain the authorization label data of the person to be entered; S52: When the data of the authorization tag of the person to enter is 0, a second warning is issued; otherwise, there is no need to issue a second warning.

10. A system for implementing a computer room safety early warning method based on digital twins as described in any one of claims 1 to 9.

Images