Artificial intelligence-based archival repository environment comprehensive evaluation method and evaluation system

Abstract

The invention discloses an archival repository environment comprehensive evaluation method and evaluation system based on artificial intelligence. The method comprises the steps that multi-dimensional environment data to be evaluated in an archival repository are acquired regularly, wherein the multi-dimensional environment data comprise a temperature value, a humidity value, an air pressure value, a safety state level and a timestamp; the method comprises the following steps: preliminarily processing multi-dimensional environment data to generate a first processing sequence and a second processing sequence, respectively extracting and obtaining a first key feature and a second key feature from the first processing sequence and the second processing sequence, importing the first processing sequence and the first key feature into a pre-constructed first calculation model, and calculating and obtaining a first evaluation coefficient; importing the second processing sequence and the second key feature into a pre-constructed second calculation model, and calculating to obtain a second evaluation coefficient; constructing a multi-dimensional feature space based on the first evaluation coefficient and the second evaluation coefficient; performing depth analysis on the multi-dimensional feature space, mutually overlapping the multi-dimensional feature space and the standard feature space, generating an evaluation factor, constructing a dynamic change curve based on the evaluation factor, comparing the dynamic change curve with a preset standard threshold curve, obtaining an abnormal time period and an abnormal area, marking an abnormal multi-dimensional feature space, and obtaining the abnormal multi-dimensional feature space. Generating an archival repository environment multidimensional thermodynamic diagram through visual analysis, and uploading and displaying the archival repository environment multidimensional thermodynamic diagram; wherein the safety state level is obtained based on judgment of a pre-established machine learning model. According to the invention, effective fusion and intelligent judgment can be carried out on the monitoring data, and abnormity identification can be carried out on the archival repository environment to carry out early warning processing.

Description

Technical Field

[0001] The present invention relates to the technical field of archive warehouse evaluation, and in particular to an artificial intelligence-based comprehensive evaluation method and evaluation system for an archive warehouse environment. Background Art

[0002] As a place for preserving and managing various important documents, the environmental conditions of archive warehouses play a vital role in the long-term preservation and effective management of archives. However, the application of environmental monitoring and assessment in archive warehouses is still in its infancy.

[0003] The existing application publication number is CN110986323A, and the document titled "A Method for Constant Temperature Control in Archive Warehouses" states that: by predicting temperature change trends, the temperature control equipment can be started and stopped in advance, thereby extending the continuous operation and continuous stop time of the temperature control equipment, thereby reducing the number of temperature control equipment state changes. The technical solution is the monitoring and alarm of a single parameter (temperature parameter), lacking comprehensive analysis and in-depth evaluation of multi-dimensional environmental data, and failing to achieve real-time intelligent optimization.

[0004] In recent years, the application of artificial intelligence in the field of archive management has made significant progress. By accurately assessing the archive environment, it can especially promptly identify the archive preservation environment, reduce the reduction or loss of archive lifespan due to environmental problems, and intervene through intelligent prediction and automatic response systems. Automatic response systems usually use a variety of monitoring devices (such as temperature and humidity sensors and smoke detectors) for measurement. However, the monitoring devices work independently and lack effective integration and intelligent judgment of monitoring data. In addition, most evaluation systems cannot provide early warnings before adverse environmental conditions occur, lacking sufficient foresight and intelligence.

[0005] Based on the above documents and existing technologies, an artificial intelligence-based comprehensive evaluation method and evaluation system for archive warehouse environment are proposed. Summary of the Invention

[0006] (1) Technical problems solved

[0007] In response to the shortcomings of the existing technology, the present invention provides an artificial intelligence-based comprehensive evaluation method and evaluation system for archive warehouse environment. First, a machine learning model is built to judge and analyze the status data, and multi-dimensional environmental data is formed by introducing relevant data such as temperature, humidity and air pressure values, so as to effectively monitor the archive warehouse in an all-round way; by importing the first processing sequence and the first key feature into a pre-built first calculation model, a first evaluation coefficient is calculated and obtained; by importing the second processing sequence and the second key feature into a pre-built second calculation model, a second evaluation coefficient is calculated and obtained, a multi-dimensional feature space is formed, and it is overlapped with the standard feature space to generate an evaluation factor, thereby constructing a dynamic change curve, obtaining abnormal time periods and abnormal areas, and marking the abnormal multi-dimensional feature space to generate a multi-dimensional heat map of the archive warehouse environment. To a certain extent, the accuracy of the model is improved, and the timely warning function is played, thereby solving the problems raised in the background technology.

[0008] (2) Technical solution

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0010] In a first aspect, the present application provides an artificial intelligence-based comprehensive evaluation method for archive warehouse environments, the method comprising:

[0011] Regularly obtain multi-dimensional environmental data to be evaluated in the archive warehouse. The multi-dimensional environmental data includes temperature, humidity, air pressure, safety status level, and timestamp. The safety status level is determined based on a pre-established machine learning model.

[0012] Preliminary processing is performed on the multidimensional environmental data to generate a first processing sequence and a second processing sequence, and a first key feature and a second key feature are extracted from the first processing sequence and the second processing sequence, respectively. The first processing sequence and the first key feature are introduced into a pre-constructed first calculation model to calculate a first evaluation coefficient; the second processing sequence and the second key feature are introduced into a pre-constructed second calculation model to calculate a second evaluation coefficient; and a multidimensional feature space is constructed based on the first evaluation coefficient and the second evaluation coefficient.

[0013] An in-depth analysis of the multidimensional feature space is conducted, and the multidimensional feature space and the standard feature space are overlapped to generate evaluation factors. A dynamic change curve is constructed based on the evaluation factors. The dynamic change curve is compared with the preset standard threshold curve to obtain abnormal time periods and abnormal areas, and the abnormal multidimensional feature space is marked. A multidimensional heat map of the archive warehouse environment is generated through visual analysis and uploaded for display.

[0014] Furthermore, the machine learning model has a first determination unit and a second determination unit built in, the first determination unit includes a first input layer, a first hidden layer, a second hidden layer and a first output layer, and the second determination unit includes a second input layer, a third hidden layer and a second output layer, and the steps of the determination process include:

[0015] First judgment unit:

[0016] Generate a first feature vector based on the state data to be determined through the first input layer;

[0017] generating a first output vector based on the first eigenvector through the first hidden layer;

[0018] generating a second output vector based on the first output vector through a second hidden layer;

[0019] Generate a first determination result based on the second output vector by the first output layer, the first determination result including a first data set that meets the safety state obtained by the first determination unit, a second data set that does not meet the safety state, and a third data set that cannot be determined whether it meets the safety state;

[0020] Second judgment unit:

[0021] generating a second feature vector based on the state data to be determined through the second input layer;

[0022] generating a third output vector based on the second eigenvector through a third hidden layer;

[0023] generating, by the third output layer, a second determination result based on the third output vector, the second determination result including a first data set that meets the safety state and a second data set that does not meet the safety state obtained by the second determination unit;

[0024] At the same time, a first state identification is given that the safety state is met, a second state identification is given that the safety state is not met, and a third state identification is given that it is impossible to judge whether the safety state is met;

[0025] The status data includes at least the collected fire water tank water level and sprinkler network pressure;

[0026] The machine learning model is trained based on the first determination result and the second determination result to obtain an updated machine learning model for the next determination analysis.

[0027] Furthermore, the machine learning model is pre-trained at least twice, and the trained machine learning model is used to perform hierarchical determination on the status data to be determined, so as to obtain the security status level corresponding to the archive warehouse, including:

[0028] First-level determination: obtain data in the first data set that does not exist in the third data set, thereby obtaining a fourth data set; obtain data in the third data set that does not exist in the first data set, thereby obtaining a fifth data set; import the fourth data set and the first state identifier, and the fifth data set and the second state identifier as input information into a machine learning model for training, thereby obtaining an updated first training model; input the state data into the first training model, perform a first-level prediction, and obtain a first prediction coefficient;

[0029] Secondary determination: obtaining data that does not exist in the second data set from the third data set to obtain a sixth data set; obtaining data that does not exist in the third data set from the second data set to obtain a seventh data set; importing the sixth data set and the first state identifier, and the seventh data set and the second state identifier as input information into a machine learning model for training to obtain an updated second training model; inputting the state data into the second training model to perform a second-level prediction to obtain a second prediction coefficient;

[0030] Comprehensive judgment: Based on the combination of the first prediction coefficient and the second prediction coefficient, a comprehensive prediction coefficient is generated, and the comprehensive prediction coefficient is compared and analyzed with the standard prediction interval [g1, g2] to obtain the corresponding safety status level;

[0031] The formula on which the comprehensive prediction coefficient is based is:

[0032] Where forecast represents the comprehensive forecast coefficient, c1 represents the first forecast coefficient, c2 represents the second forecast coefficient, θ1 and θ2 are weight coefficients, and both θ1 and θ2 are greater than 0.

[0033] Furthermore, the calculation model of the first determination unit is:

[0034]

[0035] F1=sigmoid(S1⊙w2)+b2;

[0036] F2=sigmoid(S2⊙w3)+b3;

[0037] In the formula, H1 represents the first judgment result, F1 represents the first intermediate vector, F2 represents the second intermediate vector, S1 represents the first output vector, S2 represents the second output vector; softmax() represents the normalized exponential activation function, sigmoid() represents the nonlinear activation function; w1 represents the first weight parameter, w2 represents the second weight parameter, w3 represents the third weight parameter; b1 represents the first bias parameter, b2 represents the second bias parameter, b3 represents the third bias parameter; ⊙ represents the multiplication operation, Represents the complexity measurement operation, Θ represents the complexity, θ represents the range of complexity, and θ takes the value of [Θ / 2, Θ].

[0038] Furthermore, the calculation model of the second determination unit is:

[0039] H2=softmax(G1⊙F3)*w4+b4;

[0040] F3=2tanh(w5*2S3+b5);

[0041] G1=[G2*F3+(1-G2)*G3]*w6;

[0042] In the formula, H2 represents the second judgment result, S3 represents the third output vector, and F3 represents the third intermediate vector; w4 represents the fourth weight parameter, w5 represents the fifth weight parameter, and w6 represents the sixth weight parameter; b4 represents the fourth bias parameter, and b5 represents the fifth bias parameter; G1 represents the first gating value, G2 represents the second gating value, and G3 represents the third gating value; tanh() represents the nonlinear activation function.

[0043] Furthermore, the preliminary processing of the multi-dimensional environmental data includes:

[0044] Primary processing: removing noise from the multidimensional environmental data through a bandpass filter to obtain a first processing sequence, and extracting a first key feature based on the first processing sequence, including the mean and variance;

[0045] Importing the first calculation model: constructing a first parameter matrix based on the first processing sequence, calculating a mean factor based on the average value of each column vector of the first parameter matrix, calculating a difference factor based on the variance of each column vector, and combining the mean factor and the difference factor to obtain a first evaluation coefficient of the column vector;

[0046] Secondary processing: The first processed sequence is subjected to secondary noise reduction through fast Fourier transform, and the inverse Fourier transform is used to restore it to the time domain to obtain the second processed sequence. Based on the second processed sequence, the second key features are extracted, including difference, standard deviation and interquartile range;

[0047] Importing the second calculation model: constructing a second parameter matrix based on the second processing sequence, calculating the growth factor based on the difference between adjacent vectors in each column of the second parameter matrix combined with the standard deviation, calculating the fluctuation factor based on the interquartile range of each column vector, and combining the growth factor and the fluctuation factor to obtain the second evaluation coefficient of the column vector;

[0048] Comprehensive processing: With the time acquisition period as the X-axis, the first evaluation coefficient as the Y-axis, and the second evaluation coefficient as the Z-axis, a corresponding multidimensional feature space is constructed, and the multidimensional feature space is expressed in a three-dimensional graph form.

[0049] Furthermore, the in-depth analysis of the multidimensional feature space, overlapping the multidimensional feature space and the standard feature space, generating evaluation factors, constructing a dynamic change curve based on the evaluation factors, comparing the dynamic change curve with a preset standard threshold curve, and obtaining abnormal time periods and abnormal areas, includes:

[0050] Spatial overlap: Preset a standard feature space, overlap the multidimensional feature space with the standard feature space, obtain the horizontal overlap and vertical overlap, compare the horizontal overlap and vertical overlap to obtain the overlapping area and non-overlapping area; generate an evaluation factor based on the combination of the overlapping area and the non-overlapping area;

[0051] Curve comparison: Establish a two-dimensional rectangular coordinate system and draw a dynamic change curve of the evaluation factor-time; at the same time, draw a standard threshold curve in the two-dimensional rectangular coordinate system, compare and analyze the dynamic change curve and the standard threshold curve, count the time periods when the dynamic change curve is above the standard threshold curve, and the area enclosed by the standard threshold curve, and mark them as abnormal time periods and abnormal areas.

[0052] In a second aspect, the present application provides an artificial intelligence-based comprehensive evaluation system for archive warehouse environments, the system comprising:

[0053] The data acquisition module is used to regularly obtain the multi-dimensional environmental data to be evaluated in the archive warehouse. The multi-dimensional environmental data includes temperature, humidity, air pressure, safety status level, and timestamp. The safety status level is determined based on a pre-established machine learning model.

[0054] a preliminary processing module for preliminarily processing the multidimensional environmental data, generating a first processing sequence and a second processing sequence, extracting a first key feature and a second key feature from the first processing sequence and the second processing sequence, respectively, importing the first processing sequence and the first key feature into a pre-constructed first calculation model to calculate a first evaluation coefficient; importing the second processing sequence and the second key feature into a pre-constructed second calculation model to calculate a second evaluation coefficient; and constructing a multidimensional feature space based on the first evaluation coefficient and the second evaluation coefficient;

[0055] The deep analysis module overlaps the multi-dimensional feature space and the standard feature space to generate evaluation factors, constructs a dynamic change curve based on the evaluation factors, compares the dynamic change curve with the preset standard threshold curve, obtains abnormal time periods and abnormal areas, marks the abnormal multi-dimensional feature space, and generates a multi-dimensional heat map of the archive warehouse environment through visual analysis and uploads it for display.

[0056] (3) Beneficial effects

[0057] The present invention provides an artificial intelligence-based comprehensive evaluation method and system for archive warehouse environments, which has the following beneficial effects:

[0058] 1. The present invention first trains data such as the water level of the fire water tank and the pressure of the sprinkler network by building a machine learning model. A first determination unit and a second determination unit are provided to respectively identify the first determination result and the second determination result. The trained machine learning model is used to perform a hierarchical determination on the status data to be determined, so as to obtain the corresponding safety status level of the archive warehouse. To a certain extent, the accuracy of the model is improved. Further, relevant data such as temperature, humidity and air pressure are introduced to form multi-dimensional environmental data, which effectively and comprehensively monitors the archive warehouse.

[0059] 2. The present invention obtains the first evaluation coefficient and the second evaluation coefficient by calculating the pre-built first calculation model and the second calculation model to construct a multi-dimensional feature space, and overlaps it with the preset standard feature space, defines the offset factor based on the horizontal overlap amount and the vertical overlap amount, and sets the formula to generate the evaluation factor; through curve comparison, the abnormal time period and abnormal area are identified, the abnormal multi-dimensional feature space is identified, and the abnormal changes in the archive warehouse are accurately tracked, thereby improving the accuracy of subsequent abnormal analysis and facilitating timely early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 is a flow chart of a comprehensive evaluation method for an archive warehouse environment according to an exemplary embodiment;

[0061] Figure 2 The figure is a module diagram of a comprehensive evaluation system for archive warehouse environment according to an exemplary embodiment. DETAILED DESCRIPTION

[0062] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0063] Example 1

[0064] The embodiment of the present invention provides an artificial intelligence-based comprehensive evaluation method for archive warehouse environment. Figure 1 This is a flow chart of a comprehensive evaluation method for archive warehouse environment according to an exemplary embodiment. Figure 1 , the method comprises the following steps:

[0065] S1, regularly obtains multi-dimensional environmental data to be evaluated in the archive warehouse. The multi-dimensional environmental data includes temperature, humidity, air pressure, safety status level, and timestamp. The safety status level is determined based on a pre-established machine learning model.

[0066] The machine learning model has a built-in first determination unit and a second determination unit. The first determination unit includes a first input layer, a first hidden layer, a second hidden layer, and a first output layer. The second determination unit includes a second input layer, a third hidden layer, and a second output layer. The determination process includes:

[0067] First judgment unit:

[0068] Generate a first feature vector based on the state data to be determined through the first input layer; wherein the state data includes at least the collected fire water tank water level and sprinkler network pressure, and also includes the residual flow of the electrical line and the cable temperature;

[0069] generating a first output vector based on the first eigenvector through the first hidden layer;

[0070] generating a second output vector based on the first output vector through a second hidden layer;

[0071] Generate a first determination result based on the second output vector by the first output layer, the first determination result including a first data set that meets the safety state obtained by the first determination unit, a second data set that does not meet the safety state, and a third data set that cannot be determined whether it meets the safety state;

[0072] It should be noted that the above steps are usually performed with the aid of a neural network architecture. In the neural network, the input layer is responsible for receiving raw data. In this embodiment, the raw data is the state data to be determined, and the state data is usually measured with the aid of a flow sensor and a pressure sensor. The flow sensor is used to measure the water level of the fire water tank, and the pressure sensor is used to measure the pressure of the sprinkler network. The dimension of each state data is 1 (i.e., the water level of the fire water tank and the pressure of the sprinkler network each correspond to a scalar). In this embodiment, only the water level of the fire water tank and the pressure of the sprinkler network are used as examples.

[0073] Assuming the first hidden layer has 128 neurons, the dimension of the first output vector is 128. This dimension indicates that the first feature vector is input in this layer, and the features are deeply transformed and learned, which can capture more data patterns. Assuming the second hidden layer has 256 neurons, the dimension of the second output vector is 256. By setting a higher dimension, the neural network can construct more abstract and complex feature representations between the two layers, thereby improving the data modeling ability. Finally, a softmax layer is used to generate a 3D output vector representing the probability of each category (i.e., the three situations of meeting the safe state, not meeting the safe state, and whether the safe state is not determined).

[0074] Compliant with safety status: indicates that the current state data complies with all preset safety standards, rules or specifications and does not have any known security risks. Non-compliant with safety status: indicates that the current state data violates certain safety standards or specifications and may pose a security risk or potential hazard. Unable to determine: indicates that it is impossible to determine whether the current state data violates safety standards or specifications. For example, during the training process, the complexity formed by the extracted features is too high, resulting in unstable system status or incomplete data.

[0075] Then the calculation model of the first determination unit is:

[0076]

[0077] F1=sigmoid(S1⊙w2)+b2;

[0078] F2=sigmoid(S2⊙w3)+b3;

[0079] In the formula, H1 represents the first judgment result, F1 represents the first intermediate vector, F2 represents the second intermediate vector, S1 represents the first output vector, S2 represents the second output vector; softmax() represents the normalized exponential activation function, sigmoid() represents the nonlinear activation function; w1 represents the first weight parameter, w2 represents the second weight parameter, w3 represents the third weight parameter; b1 represents the first bias parameter, b2 represents the second bias parameter, b3 represents the third bias parameter; ⊙ represents the multiplication operation, It represents the complexity measurement operation, which does not directly affect the specific value of each feature or data, but exists only as a measure. Θ represents the complexity, θ represents the range of complexity, and θ takes the value of [Θ / 2, Θ]. The operation reflects the dynamic interaction between F1 and F2, and can be nonlinearly compressed or expanded accordingly to adaptively select the weights or adjustment factors of the calculation process;

[0080] It should be noted that the first and second hidden layers are usually the core parts of the neural network for processing data, responsible for extracting features or learning certain patterns from the input information. Through these layer transformations, the model complexity gradually increases, and more nonlinear relationships can be captured, improving the data modeling ability, so that the input data will be mapped to a new representation, which is the first output vector.

[0081] In this embodiment, the first intermediate vector and the second intermediate vector are vectors generated in two intermediate steps in the above process. Their complexity can be quantified by information entropy. The general idea is to convert the elements or certain features of the vector into a probability distribution and then calculate the entropy value of the distribution.

[0082] Specifically include:

[0083] Discretize the first intermediate vector and the second intermediate vector, compress each element or feature in the vector into several intervals [0, 1] to form discrete categories, and calculate the frequency of elements in each interval;

[0084] For example:

[0085] Data 1: Assume there is a vector: A1 = [1.2, 2.4, 3.1, 4.2, 2.3, 3.5, 2.9];

[0086] The normalized vector is: [0.12, 0.24, 0.31, 0.42, 0.23, 0.35, 0.29];

[0087] Map it to different categories, assuming the intervals are:

[0088] Interval 1: [0.10, 0.20): Category A; Interval 2: [0.20, 0.30): Category B;

[0089] Interval 3: [0.30, 0.40): Category C; Interval 4: [0.40, 0.50): Category D;

[0090] Interval 5: [0.50, 0.60): Category E; Interval 6: [0.60, 0.70): Category F;

[0091] Interval 7: [0.70, 0.80): Category G; Interval 8: [0.80, 0.90): Category H;

[0092] Interval 9: [0.90, 1.00]: Category I;

[0093] The discretized vector is: [A, B, C, D, B, C, B];

[0094] Then the information entropy is: -((1 / 7)log2(1 / 7)+(3 / 7)log2(3 / 7)+(2 / 7)log2(2 / 7)+(1 / 7)log2(1 / 7))≈1.842;

[0095] Data 2: Assume there is a vector: A2 = [1.2, 3.1, 2.8, 3.4, 2.2, 2.8, 1.2];

[0096] The normalized vector is: [0.12, 0.31, 0.28, 0.34, 0.22, 0.28, 0.12];

[0097] Based on the interval category of data 1, the discretized vector is: [A, C, B, C, B, B, A];

[0098] Then the information entropy is: -((2 / 7)log2(2 / 7)+(3 / 7)log2(3 / 7)+(2 / 7)log2(2 / 7)+0)≈1.5575;

[0099] If data 1 is greater than data 2, it means that the vector corresponding to data 1 has higher information content and is more complex;

[0100] Second judgment unit:

[0101] generating a second feature vector based on the state data to be determined through the second input layer;

[0102] generating a third output vector based on the second eigenvector through a third hidden layer;

[0103] generating, by the third output layer, a second determination result based on the third output vector, the second determination result including a first data set that meets the safety state and a second data set that does not meet the safety state obtained by the second determination unit;

[0104] At the same time, a first state identification is given that the safety state is met, a second state identification is given that the safety state is not met, and a third state identification is given that it is impossible to judge whether the safety state is met;

[0105] It should be noted that the first status identifier, the second status identifier, and the third status identifier are combined with the corresponding information and processed to generate a hash value, which can be stored in a database, file, or other storage medium and can be used for subsequent data verification, search, and other operations. The specific steps involved are not described in detail here;

[0106] Then the calculation model of the second determination unit is:

[0107] H2=sofmtax(G1⊙F3)*w4+b4;

[0108] F3=2tanh(w5*2S3+b5);

[0109] G1=[G3*F3+(-G2)*G3]*w6;

[0110] Wherein, H2 represents the second determination result, S3 represents the third output vector, F3 represents the third intermediate vector; w4 represents the fourth weight parameter, w5 represents the fifth weight parameter, w6 represents the sixth weight parameter; b4 represents the fourth bias parameter, b5 represents the fifth bias parameter; G1 represents the first gating value, G2 represents the second gating value, G3 represents the third gating value; tanh() represents the nonlinear activation function;

[0111] It should be noted that, under normal circumstances, the gate value is calculated based on certain activation functions (such as sigmoid or tanh), and their values will be between 0 and 1, which is used to "control" the information flow of other tensors; the complexity calculation of defining the second determination unit is much smaller than the complexity of the first determination unit. The features obtained in the second determination unit process can be regarded as shallow features, while the features obtained in the first determination unit can be regarded as deep features; in the case of a deep level (too high complexity), it is easy to increase the difficulty of training. In this embodiment, it will be impossible to determine whether it meets the safety state. Therefore, this application uses a shallow level (lower complexity) to integrate training with a deep level to ensure the accuracy of the model, which is conducive to subsequent training optimization.

[0112] The machine learning model is trained based on the first determination result and the second determination result to obtain an updated machine learning model for the next determination analysis; at least two levels of pre-trained machine learning models are used to perform hierarchical determination on the status data to be determined to obtain the corresponding security status level of the archive warehouse, including:

[0113] First-level determination: obtain data in the first data set that does not exist in the third data set, thereby obtaining a fourth data set; obtain data in the third data set that does not exist in the first data set, thereby obtaining a fifth data set; import the fourth data set and the first state identifier, and the fifth data set and the second state identifier as input information into a machine learning model for training, thereby obtaining an updated first training model; input the state data into the first training model, perform a first-level prediction, and obtain a first prediction coefficient;

[0114] Secondary determination: obtaining data that does not exist in the second data set from the third data set to obtain a sixth data set; obtaining data that does not exist in the third data set from the second data set to obtain a seventh data set; importing the sixth data set and the first state identifier, and the seventh data set and the second state identifier as input information into a machine learning model for training to obtain an updated second training model; inputting the state data into the second training model to perform a second-level prediction to obtain a second prediction coefficient;

[0115] It should be noted that by inputting the state data to be determined into the first training model or the second training model, the prediction result obtained will undergo a conversion analysis, the purpose of which is to convert the output result of the first training model or the second training model into a predicted value; this can be processed by weighted averaging, weighted voting, etc., which will not be described in detail here;

[0116] Comprehensive judgment: Based on the combination of the first prediction coefficient and the second prediction coefficient, a comprehensive prediction coefficient is generated. The formula based on the comprehensive prediction coefficient is:

[0117] Where, forecast represents the comprehensive forecast coefficient, c1 represents the first forecast coefficient, c2 represents the second forecast coefficient, θ1 and θ2 are weight coefficients, and both θ1 and θ2 are greater than 0;

[0118] Compare and analyze the comprehensive prediction coefficient with the standard prediction interval [g1, g2]:

[0119] When forecast < g1, a class level identifier is assigned, and this value is combined with the class level identifier to generate a class safety level, and the safety level value of the environment at that time is judged to be low;

[0120] When g1≤forecast≤g2, a second-class level identification is assigned, and this value is combined with the second-class level identification to generate a first-class safety level, and the safety level of the environment at that time is determined to be average;

[0121] When forecast>g2, a three-level grade identifier is assigned, and this value is combined with the three-level grade identifier to generate a three-level security level, and the security level value of the environment at that time is determined to be higher;

[0122] S2, preliminarily processing the multidimensional environmental data to generate a first processing sequence and a second processing sequence, extracting a first key feature and a second key feature from the first processing sequence and the second processing sequence, respectively, importing the first processing sequence and the first key feature into a pre-constructed first calculation model to calculate a first evaluation coefficient; importing the second processing sequence and the second key feature into a pre-constructed second calculation model to calculate a second evaluation coefficient; and constructing a multidimensional feature space based on the first evaluation coefficient and the second evaluation coefficient;

[0123] Preliminary processing of multi-dimensional environmental data, including:

[0124] Primary processing: removing noise from the multidimensional environmental data through a bandpass filter to obtain a first processing sequence, and extracting a first key feature based on the first processing sequence, including the mean and variance;

[0125] Import the first calculation model:

[0126] Construct the first parameter matrix based on the first processing sequence:

[0127] Where Ja1 represents the first parameter matrix, j represents the first processing sequence, and j 1,1 、j 1,2 、j 1,3 and j 1,4 Respectively represent the temperature value, humidity value, air pressure value and safety status level at the first moment, j 2,1 、j 2,2 、j 2,3 and j 2,4 Respectively represent the temperature value, humidity value, air pressure value and safety status level at the second moment, j N0,1 、j N0,2 、j N0,3 and j N0,4 Respectively represent the temperature value, humidity value, air pressure value and safety status level at the N0th moment;

[0128] A mean factor is calculated based on the average value of each column vector of the first parameter matrix, a difference factor is calculated based on the variance of each column vector, and the mean factor and the difference factor are combined to obtain a first evaluation coefficient of the column vector;

[0129] The formula for calculating the first evaluation coefficient is:

[0130]

[0131] In the formula, assess1 represents the first assessment coefficient, μ represents the mean factor, σ represents the difference factor, and α1 and α1 are weight coefficients;

[0132] Among them, take the first group of column vectors as an example:

[0133] The formula for calculating the mean factor is:

[0134] The formula for calculating the difference factor is:

[0135] Where m represents the number of column vectors, and its value range is [1, N0];

[0136] Secondary processing: The first processed sequence is subjected to secondary noise reduction through fast Fourier transform, and the inverse Fourier transform is used to restore it to the time domain to obtain the second processed sequence. Based on the second processed sequence, the second key features are extracted, including difference, standard deviation and interquartile range;

[0137] Import the second calculation model:

[0138] Construct the second parameter matrix based on the second processing sequence:

[0139] Where Ja2 represents the second parameter matrix, y represents the second processing sequence, each row vector represents the second processing sequence obtained under N0 time measurement cycles, and each column vector represents the temperature value, humidity value, air pressure value, and safety status level corresponding to the second processing sequence;

[0140] An increase factor is calculated based on the difference between adjacent vectors in each column of the second parameter matrix and combined with the standard deviation. A fluctuation factor is calculated based on the interquartile range of each column vector. The increase factor and the fluctuation factor are combined to obtain a second evaluation coefficient of the column vector.

[0141] The formula for calculating the second evaluation coefficient is:

[0142]

[0143] In the formula, assess2 represents the second assessment coefficient, rate represents the increase rate factor, fluct represents the fluctuation factor, and β1 and β2 are both weight coefficients;

[0144] Among them, take the first group of column vectors as an example:

[0145] The formula for calculating the growth factor is:

[0146] The formula for calculating the volatility factor is:

[0147] Where, IOR represents the interquartile range, which is calculated as the difference between the third quartile and the first quartile and is used to measure the dispersion or volatility of the data; Represents the average value of the first set of vectors, y n,1 represents a column vector;

[0148] Comprehensive processing: With the time measurement period N0 as the X-axis, the first evaluation coefficient assess1 as the Y-axis, and the second evaluation coefficient assess2 as the Z-axis, a corresponding multidimensional feature space is constructed, and the multidimensional feature space is expressed in a three-dimensional graph form;

[0149] It should be noted that by combining the temperature value, humidity value, air pressure value and safety status level for analysis, the precision and accuracy of the system analysis can be further improved in the judgment process;

[0150] S3: Conduct an in-depth analysis of the multidimensional feature space, overlap the multidimensional feature space with the standard feature space, generate evaluation factors, construct a dynamic change curve based on the evaluation factors, compare the dynamic change curve with the preset standard threshold curve, obtain abnormal time periods and abnormal areas, mark the abnormal multidimensional feature space, and generate a multidimensional heat map of the archive warehouse environment through visual analysis and upload it for display;

[0151] In-depth analysis of the multidimensional feature space is performed, and the multidimensional feature space and the standard feature space are overlapped to generate evaluation factors. Based on the evaluation factors, a dynamic change curve is constructed. The dynamic change curve is compared with the preset standard threshold curve to obtain the abnormal period and abnormal area, including:

[0152] Spatial overlap: Preset a standard feature space, overlap the multidimensional feature space with the standard feature space, obtain the horizontal overlap and vertical overlap, compare the horizontal overlap and vertical overlap to obtain the overlapping area and non-overlapping area; generate an evaluation factor based on the combination of the overlapping area and the non-overlapping area;

[0153] The formula for calculating the evaluation factor is:

[0154] Where, Evaluation represents the evaluation factor, fmj represents the non-overlapping area, cmj represents the overlapping area, and λ represents the offset factor;

[0155] The calculation formula of the offset factor is:

[0156] Where hx represents the horizontal overlap, zx represents the vertical overlap, bz represents the multidimensional feature space, and Bz represents the standard feature space;

[0157] Curve comparison: Establish a two-dimensional rectangular coordinate system and draw a dynamic change curve of the evaluation factor-time. At the same time, draw a standard threshold curve in the two-dimensional rectangular coordinate system, compare and analyze the dynamic change curve and the standard threshold curve, and count the time periods when the dynamic change curve is above the standard threshold curve and the area enclosed by the standard threshold curve, and mark them as abnormal time periods and abnormal areas.

[0158] It should be noted that the abnormal period indicates the time period when the archive warehouse environment is abnormal during the evaluation process, and the abnormal area reflects the severity of the archive warehouse environment during the evaluation process. The larger the abnormal area, the worse the environment and the greater the risk of fire. By targeted analysis of the abnormal time period and abnormal area, anomalies are marked in the multidimensional feature space, and a multidimensional heat map of the archive warehouse environment is drawn through visual analysis to accurately track abnormal changes in the archive warehouse and improve the accuracy of subsequent anomaly analysis. Based on the multidimensional heat map of the archive warehouse environment, a clustering algorithm (such as K-means or DBSCAN) can be used to group the abnormal points in the map, and by analyzing which areas in the heat map are most concentrated with abnormalities, combined with the sensors or equipment involved in the system, the pattern or cause of the abnormality can be identified. The specific process will not be described in detail here.

[0159] In summary of the above technical solutions: the present invention first conducts judgment and analysis on the status data by building a machine learning model, and forms multi-dimensional environmental data by introducing relevant data such as temperature values, humidity values and air pressure values, so as to effectively monitor the archive warehouse in an all-round way; by importing the first processing sequence and the first key feature into a pre-built first calculation model, the first evaluation coefficient is calculated and obtained, and by importing the second processing sequence and the second key feature into a pre-built second calculation model, the second evaluation coefficient is calculated and obtained to form a multi-dimensional feature space, and overlap it with the standard feature space to generate an evaluation factor, thereby constructing a dynamic change curve, obtaining abnormal time periods and abnormal areas, and marking the abnormal multi-dimensional feature space to generate a multi-dimensional heat map of the archive warehouse environment, which improves the accuracy of the model to a certain extent and plays a role in timely early warning.

[0160] Example 2

[0161] The embodiment of the present invention provides an archive warehouse environment comprehensive evaluation system based on artificial intelligence. Figure 2 This is a module diagram of a comprehensive evaluation system for archive warehouse environment according to an exemplary embodiment. Figure 2 The system includes: a data acquisition module, a preliminary processing module and a depth analysis module, and the data acquisition module, the preliminary processing module and the depth analysis module are communicatively connected;

[0162] The data acquisition module is used to regularly obtain the multi-dimensional environmental data to be evaluated in the archive warehouse. The multi-dimensional environmental data includes temperature, humidity, air pressure, safety status level, and timestamp. The safety status level is determined based on a pre-established machine learning model.

[0163] a preliminary processing module for preliminarily processing the multidimensional environmental data, generating a first processing sequence and a second processing sequence, extracting a first key feature and a second key feature from the first processing sequence and the second processing sequence, respectively, importing the first processing sequence and the first key feature into a pre-constructed first calculation model to calculate a first evaluation coefficient; importing the second processing sequence and the second key feature into a pre-constructed second calculation model to calculate a second evaluation coefficient; and constructing a multidimensional feature space based on the first evaluation coefficient and the second evaluation coefficient;

[0164] The deep analysis module overlaps the multi-dimensional feature space and the standard feature space to generate evaluation factors, constructs a dynamic change curve based on the evaluation factors, compares the dynamic change curve with the preset standard threshold curve, obtains abnormal time periods and abnormal areas, marks the abnormal multi-dimensional feature space, and generates a multi-dimensional heat map of the archive warehouse environment through visual analysis and uploads it for display.

[0165] In this application, the weight coefficients involved are determined using the coefficient of variation method, which is a method of weighting each indicator based on the degree of variation between the current value of each evaluation indicator and the target value; if the numerical difference of a certain indicator is large, it can clearly distinguish the evaluated objects, indicating that the indicator has rich discrimination information, and thus the indicator should be given a larger weight; conversely, if the numerical difference of each evaluated object on a certain indicator is small, then the ability of this indicator to distinguish the evaluation objects is weak, and thus the indicator should be given a smaller weight; this method directly uses the information contained in each indicator to obtain the weight of the indicator through calculation, and therefore is objective.

[0166] In the application, the several formulas involved are all calculated by taking their numerical values after removing the dimensions, and the formula is a formula of the most recent real situation obtained by collecting a large amount of data and performing software simulation. The formula is set by technical personnel in this field according to actual conditions.

[0167] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0168] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, and may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment as needed.

[0169] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A comprehensive evaluation method for archive warehouse environment based on artificial intelligence, characterized by: The method comprises: Regularly obtain multi-dimensional environmental data to be evaluated in the archive warehouse. The multi-dimensional environmental data includes temperature, humidity, air pressure, safety status level, and timestamp. The safety status level is determined based on a pre-established machine learning model. Preliminary processing is performed on the multidimensional environmental data to generate a first processing sequence and a second processing sequence, and a first key feature and a second key feature are extracted from the first processing sequence and the second processing sequence, respectively. The first processing sequence and the first key feature are introduced into a pre-constructed first calculation model to calculate a first evaluation coefficient; the second processing sequence and the second key feature are introduced into a pre-constructed second calculation model to calculate a second evaluation coefficient; and a multidimensional feature space is constructed based on the first evaluation coefficient and the second evaluation coefficient. An in-depth analysis of the multidimensional feature space is conducted, and the multidimensional feature space and the standard feature space are overlapped to generate evaluation factors. A dynamic change curve is constructed based on the evaluation factors. The dynamic change curve is compared with the preset standard threshold curve to obtain abnormal time periods and abnormal areas, mark the abnormal multidimensional feature space, and generate a multidimensional heat map of the archive warehouse environment through visual analysis.

2. The method for comprehensive evaluation of archive warehouse environment based on artificial intelligence according to claim 1 is characterized in that: The machine learning model has a built-in first determination unit and a second determination unit. The first determination unit includes a first input layer, a first hidden layer, a second hidden layer, and a first output layer. The second determination unit includes a second input layer, a third hidden layer, and a second output layer. The steps of the determination process include: First judgment unit: Generate a first feature vector based on the state data to be determined through the first input layer; Generate a first output vector based on the first eigenvector through the first hidden layer; generating a second output vector based on the first output vector through a second hidden layer; Generate a first determination result based on the second output vector by the first output layer, the first determination result including a first data set that meets the safety state obtained by the first determination unit, a second data set that does not meet the safety state, and a third data set that cannot be determined whether it meets the safety state; Second judgment unit: generating a second feature vector based on the state data to be determined through the second input layer; generating a third output vector based on the second eigenvector through a third hidden layer; generating, by the third output layer, a second determination result based on the third output vector, the second determination result including a first data set that meets the safety state and a second data set that does not meet the safety state obtained by the second determination unit; At the same time, a first state identification is given that the safety state is met, a second state identification is given that the safety state is not met, and a third state identification is given that it is impossible to judge whether the safety state is met; The status data includes at least the collected fire water tank water level and sprinkler network pressure; The machine learning model is trained based on the first determination result and the second determination result to obtain an updated machine learning model for the next determination analysis.

3. The method for comprehensive evaluation of archive warehouse environment based on artificial intelligence according to claim 2 is characterized in that: The machine learning model is pre-trained at least twice, and the trained machine learning model is used to perform hierarchical judgment on the status data to be judged to obtain the security status level corresponding to the archive warehouse, including: First-level determination: obtain data in the first data set that does not exist in the third data set, thereby obtaining a fourth data set; obtain data in the third data set that does not exist in the first data set, thereby obtaining a fifth data set; import the fourth data set and the first state identifier, and the fifth data set and the second state identifier as input information into a machine learning model for training, thereby obtaining an updated first training model; input the state data into the first training model, perform a first-level prediction, and obtain a first prediction coefficient; Secondary determination: obtaining data that does not exist in the second data set from the third data set to obtain a sixth data set; obtaining data that does not exist in the third data set from the second data set to obtain a seventh data set; importing the sixth data set and the first state identifier, and the seventh data set and the second state identifier as input information into a machine learning model for training to obtain an updated second training model; inputting the state data into the second training model to perform a second-level prediction to obtain a second prediction coefficient; Comprehensive judgment: Based on the combination of the first prediction coefficient and the second prediction coefficient, a comprehensive prediction coefficient is generated, and the comprehensive prediction coefficient is compared and analyzed with the standard prediction interval [g1, g2] to obtain the corresponding safety status level; The formula for the comprehensive prediction coefficient is: Where forecast represents the comprehensive forecast coefficient, c1 represents the first forecast coefficient, c2 represents the second forecast coefficient, θ1 and θ2 are weight coefficients, and both θ1 and θ2 are greater than 0.

4. The method for comprehensive evaluation of archive warehouse environment based on artificial intelligence according to claim 2 is characterized in that: The calculation model of the first determination unit is: F1=sigmoid(S1☉w2)+b2; F2=sigmoid(S2☉w3)+b3; In the formula, H1 represents the first judgment result, F1 represents the first intermediate vector, F2 represents the second intermediate vector, S1 represents the first output vector, S2 represents the second output vector; softmax() represents the normalized exponential activation function, sigmoid() represents the nonlinear activation function; w1 represents the first weight parameter, w2 represents the second weight parameter, w3 represents the third weight parameter; b1 represents the first bias parameter, b2 represents the second bias parameter, b3 represents the third bias parameter; ⊙ represents the multiplication operation, Represents the complexity measurement operation, Θ represents the complexity, θ represents the range of complexity, and θ takes the value of [Θ / 2, Θ].

5. The method for comprehensive evaluation of archive warehouse environment based on artificial intelligence according to claim 2 is characterized in that: The calculation model of the second determination unit is: H2=softmax(G1☉F3)*w4+b4; E3=2tanh(w5**2S3+b5); G1=[G2*F3+(1-G2)*G3]*w6; Wherein, H2 represents the second determination result, S3 represents the third output vector, F3 represents the third intermediate vector; w4 represents the fourth weight parameter, w5 represents the fifth weight parameter, and w6 represents the sixth weight parameter; b4 represents the fourth bias parameter, b5 represents the fifth bias parameter; G1 represents the first gating value, G2 represents the second gating value, G3 represents the third gating value; tanh() represents a nonlinear activation function.

6. The method for comprehensive evaluation of archive warehouse environment based on artificial intelligence according to claim 1 is characterized in that: The preliminary processing of multi-dimensional environmental data includes: Primary processing: removing noise from the multidimensional environmental data through a bandpass filter to obtain a first processing sequence, and extracting a first key feature based on the first processing sequence, including the mean and variance; Importing the first calculation model: constructing a first parameter matrix based on the first processing sequence, calculating a mean factor based on the average value of each column vector of the first parameter matrix, calculating a difference factor based on the variance of each column vector, and combining the mean factor and the difference factor to obtain a first evaluation coefficient of the column vector; Secondary processing: The first processed sequence is subjected to secondary noise reduction through fast Fourier transform, and the inverse Fourier transform is used to restore it to the time domain to obtain the second processed sequence. Based on the second processed sequence, the second key features are extracted, including difference, standard deviation and interquartile range; Importing the second calculation model: constructing a second parameter matrix based on the second processing sequence, calculating the growth factor based on the difference between adjacent vectors in each column of the second parameter matrix combined with the standard deviation, calculating the fluctuation factor based on the interquartile range of each column vector, and combining the growth factor and the fluctuation factor to obtain the second evaluation coefficient of the column vector; Comprehensive processing: With the time acquisition period as the X-axis, the first evaluation coefficient as the Y-axis, and the second evaluation coefficient as the Z-axis, a corresponding multidimensional feature space is constructed, and the multidimensional feature space is expressed in a three-dimensional graph form.

7. The method for comprehensive evaluation of archive warehouse environment based on artificial intelligence according to claim 1 is characterized in that: The in-depth analysis of the multidimensional feature space, overlapping the multidimensional feature space and the standard feature space, generating evaluation factors, constructing a dynamic change curve based on the evaluation factors, comparing the dynamic change curve with a preset standard threshold curve, and obtaining abnormal time periods and abnormal areas, includes: Spatial overlap: Preset a standard feature space, overlap the multidimensional feature space with the standard feature space, obtain the horizontal overlap and vertical overlap, compare the horizontal overlap and vertical overlap to obtain the overlapping area and non-overlapping area; generate an evaluation factor based on the combination of the overlapping area and the non-overlapping area; Curve comparison: Establish a two-dimensional rectangular coordinate system and draw a dynamic change curve of the evaluation factor-time; at the same time, draw a standard threshold curve in the two-dimensional rectangular coordinate system, compare and analyze the dynamic change curve and the standard threshold curve, count the time periods when the dynamic change curve is above the standard threshold curve, and the area enclosed by the standard threshold curve, and mark them as abnormal time periods and abnormal areas.

8. The comprehensive evaluation system of archive warehouse environment based on artificial intelligence is characterized by: include: The data acquisition module is used to regularly obtain the multi-dimensional environmental data to be evaluated in the archive warehouse. The multi-dimensional environmental data includes temperature, humidity, air pressure, safety status level, and timestamp. The safety status level is determined based on a pre-established machine learning model. a preliminary processing module for preliminarily processing the multidimensional environmental data, generating a first processing sequence and a second processing sequence, extracting a first key feature and a second key feature from the first processing sequence and the second processing sequence, respectively, importing the first processing sequence and the first key feature into a pre-constructed first calculation model to calculate a first evaluation coefficient; importing the second processing sequence and the second key feature into a pre-constructed second calculation model to calculate a second evaluation coefficient; and constructing a multidimensional feature space based on the first evaluation coefficient and the second evaluation coefficient; The deep analysis module overlaps the multi-dimensional feature space and the standard feature space to generate evaluation factors, constructs a dynamic change curve based on the evaluation factors, compares the dynamic change curve with the preset standard threshold curve, obtains abnormal time periods and abnormal areas, marks the abnormal multi-dimensional feature space, and generates a multi-dimensional heat map of the archive warehouse environment through visual analysis and uploads it for display.

Images