A safety engineering optimization method and system based on artificial intelligence

Abstract

The present invention discloses a safety engineering optimization method and system based on artificial intelligence, which belongs to the technical field of safety engineering optimization. The method includes data preparation, data preprocessing, elevator operation status evaluation, elevator fault prediction and elevator safety engineering optimization. The present invention adopts a long short-term memory model with enhanced parameter optimization to evaluate the elevator operation status, increase the diversity of sparrow population, improve the global search ability of the algorithm, better optimize parameters, and improve the model prediction effect; adopts an autoencoder long short-term memory parallel model to predict elevator faults, and better captures rich feature representations by processing the autoencoder and the long short-term memory network in parallel, improves the model prediction accuracy, and performs integrated learning through gradient boosting technology to enhance the model robustness; integrates elevator operation status evaluation information and elevator fault prediction information to achieve elevator safety engineering optimization, reduce the influence of subjective factors, and effectively improve the overall stability and reliability of the elevator system.

Description

Technical Field

[0001] The present invention belongs to the technical field of safety engineering optimization, and specifically refers to a safety engineering optimization method and system based on artificial intelligence. Background Art

[0002] Safety engineering optimization uses artificial intelligence technology to conduct comprehensive analysis and processing of relevant data of the elevator system to improve the safety, reliability and operational efficiency of the elevator, thereby realizing intelligent safety management of the elevator system, reducing the accident rate, and better ensuring passenger safety. However, in the existing safety engineering optimization process, there are technical problems that the elevator operation status assessment has high requirements for model prediction accuracy and requires further parameter optimization, while the parameter optimization capability of the traditional sparrow search algorithm is limited, resulting in poor model prediction effect and affecting the reliability of the elevator operation status assessment; there are technical problems that the elevator fault data is complex and often contains noise, and the prediction accuracy and robustness of a single model in elevator fault prediction are poor, which is easy to cause false alarms and increase unnecessary maintenance costs; there are technical problems that the traditional elevator safety engineering optimization relies on the experience and judgment of professional technicians, is easily affected by subjective factors, and requires a lot of time and human resources. The process is time-consuming and labor-intensive, resulting in low efficiency and poor effectiveness of elevator safety engineering optimization. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a safety engineering optimization method and system based on artificial intelligence. In view of the fact that in the existing safety engineering optimization process, the elevator operation status assessment has high requirements for model prediction accuracy and needs further parameter optimization, while the parameter optimization capability of the traditional sparrow search algorithm is limited, resulting in poor model prediction effect and affecting the reliability of the elevator operation status assessment. The present invention creatively adopts a long short-term memory model with enhanced parameter optimization to evaluate the elevator operation status. By introducing tent chaos mapping, adaptive inertia weights and random inertia weights, the diversity of the sparrow population is increased, and it is effectively avoided to fall into the local optimal solution, thereby improving the global search capability of the algorithm, better optimizing parameters, and improving the model prediction effect, thereby enhancing the reliability of the elevator operation status assessment. In view of the fact that in the existing safety engineering optimization process, the elevator fault data is complex and often carries noise, and the single model is ineffective in predicting elevator faults. In order to solve the technical problems of poor measurement accuracy and robustness, easy to cause false alarms and increase unnecessary maintenance costs, this solution creatively adopts the autoencoder long short-term memory parallel model to predict elevator faults. By processing the autoencoder and the long short-term memory network in parallel, it can better capture rich feature representations and improve the prediction accuracy of the model. Through integrated learning with gradient boosting technology, the robustness of the model is enhanced, thereby reducing the false alarm rate and maintenance costs. In order to solve the technical problems of low efficiency and poor effectiveness of elevator safety engineering optimization, traditional elevator safety engineering optimization relies on the experience and judgment of professional technicians, is easily affected by subjective factors, and requires a lot of time and human resources. This process is time-consuming and labor-intensive. This solution creatively integrates elevator operation status evaluation information and elevator fault prediction information to achieve elevator safety engineering optimization, reduce the influence of subjective factors, effectively improve the overall stability and reliability of the elevator system, and promote the intelligent development of elevator system management.

[0004] The technical solution adopted by the present invention is as follows: The present invention provides a safety engineering optimization method based on artificial intelligence, which comprises the following steps:

[0005] Step S1: data preparation;

[0006] Step S2: data preprocessing;

[0007] Step S3: Elevator operation status evaluation;

[0008] Step S4: Elevator failure prediction;

[0009] Step S5: Elevator safety engineering optimization.

[0010] Furthermore, in step S1, the data preparation is used to collect original data required for elevator safety engineering optimization, specifically by collecting elevator system related data to obtain elevator safety original data, and the elevator safety original data includes elevator operation data, environmental data, historical fault data and maintenance records.

[0011] Further, in step S2, the data preprocessing is used to preprocess the elevator safety original data, specifically, by preprocessing the elevator safety original data to obtain elevator status integrated data and elevator fault integrated data, including the following steps:

[0012] Step S21: data cleaning, specifically, cleaning the original elevator safety data by deleting outliers, filling missing values, and removing duplicate values ​​to obtain elevator safety cleansing data;

[0013] Step S22: normalization processing, specifically, performing data normalization processing on the elevator safety cleaning data to obtain elevator safety standard data;

[0014] Step S23: Data integration, specifically, integrating the elevator operation data and environmental data in the elevator safety standard data to obtain elevator status integrated data, and integrating the historical fault data and maintenance records in the elevator safety standard data to obtain elevator fault integrated data.

[0015] Further, in step S3, the elevator operation status evaluation is used to evaluate the current operation status of the elevator, specifically, based on the elevator state integration data, using the long short-term memory model with enhanced parameter optimization to evaluate the elevator operation status, and obtain the elevator operation status evaluation information, including steps S31-S34:

[0016] Step S31: constructing a long short-term memory model, specifically constructing a long short-term memory model including an input gate, a forget gate, a memory unit and an output gate to obtain a long short-term memory model;

[0017] Step S32: improving the sparrow search algorithm, specifically, introducing tent chaos mapping, adaptive inertia weight and random inertia weight into the sparrow search algorithm, improving the sparrow search algorithm, and obtaining an improved sparrow search algorithm;

[0018] The improved sparrow search algorithm comprises the following steps:

[0019] Step S321: Initializing the sparrow population through tent chaos mapping to increase the diversity of initial solutions, including steps S3211-S3212:

[0020] Step S3211: Randomly generate each sparrow position in the interval [0,1], and the calculation formula is:

[0021] ;

[0022] In the formula, x ij is the position of the i-th sparrow in the j-th dimension, i is the sparrow index, j is the solution space dimension index, d is the matrix dimension, which is used to represent the solution space dimension, which is equal to the total number of model parameters that need to be optimized, rand(0,1) is a random number generated in the interval [0,1], pos i is the i-th sparrow position, x i1 is the position of the i-th sparrow in the first dimension, x i2 is the position of the i-th sparrow in the second dimension, x id is the position of the i-th sparrow in the d-th dimension;

[0023] Step S3212: Perform tent chaos mapping on each sparrow position until the maximum mapping times are reached to obtain the initial population of sparrows;

[0024] The calculation formula of the tent chaos map is:

[0025] ;

[0026] Where k is the mapping index, is the position of the i-th sparrow in the j-th dimension of the k+1-th mapping, is the position of the i-th sparrow in the k-th mapping on the j-th dimension;

[0027] Step S322: construct a fitness function to evaluate the quality of each sparrow position, specifically, use the cross entropy loss function as the fitness function, and calculate the fitness value of each sparrow position through the fitness function;

[0028] Step S323: updating the position of the finder, specifically introducing an adaptive inertia weight in the updating of the position of the finder, and updating the position of the finder. The finder specifically refers to a sparrow looking for food. The calculation formula for updating the position of the finder is:

[0029] ;

[0030] In the formula, is the adaptive inertia weight, exp(·) is the natural exponential function, t is the iteration index, and t max is the maximum number of iterations, is the position of the i-th sparrow at the t+1th iteration, is the position of the i-th sparrow at the t-th iteration, is an adjustment parameter with a value range of [0, 1], used to adjust the amplitude of position update, R1 is a warning value with a value range of [0, 1], val is a safety value with a value range of [0.5, 1], R2 is a random number that follows a normal distribution, L is a 1×d matrix with all elements being 1, d is the matrix dimension, used to represent the dimension of the solution space;

[0031] Step S324: Follower position update, specifically introducing random inertia weight in follower position update, and performing follower position update. The follower specifically refers to a sparrow that follows the finder to find food. The calculation formula for the follower position update is:

[0032] ;

[0033] In the formula, is the random inertia weight, is the random inertia weight mean, is the standard deviation of the random inertia weight, which is used to control the fluctuation range of the random inertia weight. N (0, 1) is a random number that obeys the standard normal distribution. is the best position occupied by the finder at the t+1th iteration, is the global worst position at the tth iteration, A is a 1×d matrix with values ​​of {-1, 1}, and T is the transpose operation;

[0034] Step S325: The position of the alerter is updated. The alerter specifically refers to the sparrow responsible for identifying potential dangers. The calculation formula is:

[0035] ;

[0036] In the formula, is the global optimal position at the tth iteration, f i is the fitness value of the i-th sparrow position, f g is the fitness value of the global optimal position, f w is the fitness value of the global worst position, is a very small constant used to avoid the situation where the denominator is zero, and R3 is a random number uniformly distributed in the interval [-1, 1];

[0037] Step S326: updating the position of the sparrow by repeatedly executing steps S323 to S325 until the maximum number of iterations is reached, and calculating the global optimal position as the optimal parameter of the model by iteration, wherein the maximum number of iterations is 500;

[0038] Step S33: constructing a long short-term memory model with enhanced parameter optimization, specifically by training the long short-term memory model and using an improved sparrow search algorithm to perform parameter tuning to obtain a long short-term memory model with enhanced parameter optimization;

[0039] Step S34: Calculate the elevator operation status evaluation result, specifically, evaluate the elevator operation status based on the elevator status integration data through the long short-term memory model with enhanced parameter optimization to obtain the elevator operation status evaluation information.

[0040] Further, in step S4, the elevator fault prediction is used to predict the probability of future elevator failures, specifically, based on the elevator fault integration data and the elevator operation status evaluation information, the autoencoder long short-term memory parallel model is used to perform elevator fault prediction to obtain elevator fault prediction information;

[0041] The autoencoder long short-term memory parallel model includes an input layer, a parallel autoencoder long short-term memory layer, a gradient boosting prediction layer and an output layer;

[0042] The parallel autoencoder long short-term memory layer includes a parallel autoencoder subnet and a long short-term memory subnet;

[0043] The autoencoder subnet is used to learn data features in an unsupervised manner;

[0044] The long short-term memory subnetwork is used to process time series data and capture long-term dependencies;

[0045] The gradient boosting prediction layer is used to comprehensively process the output of the long short-term memory layer of the parallel autoencoder and generate a prediction result;

[0046] The elevator failure prediction comprises the following steps:

[0047] Step S41: constructing an input layer for receiving input data, specifically receiving elevator fault integration data and elevator operation status evaluation information as input data of the model through the input layer;

[0048] Step S42: Constructing a parallel autoencoder long short-term memory layer, including steps S421-S422:

[0049] Step S421: construct an autoencoder subnet, specifically construct an autoencoder subnet including an encoder, a repeated vector layer, a decoder, and a reconstruction loss layer, and calculate the input data through the autoencoder subnet to obtain the output features of the autoencoder subnet;

[0050] Step S422: constructing a long short-term memory subnet, specifically, using a long short-term memory model including an input gate, a forget gate, a memory unit, and an output gate as a long short-term memory subnet, calculating the input data through the long short-term memory subnet, and obtaining the long short-term memory subnet output feature;

[0051] Step S43: constructing a gradient boosting prediction layer, specifically using a lightweight gradient boosting model as the gradient boosting prediction layer, and integrating the output features of the autoencoder subnet and the output features of the long short-term memory subnet through the gradient boosting prediction layer to obtain the final prediction result;

[0052] Step S44: construct an output layer for outputting the final prediction result;

[0053] Step S45: constructing an autoencoder long short-term memory parallel model, specifically, constructing the autoencoder long short-term memory parallel model through the input layer of step S41, the parallel autoencoder long short-term memory layer of step S42, the gradient boosting prediction layer of step S43 and the output layer of step S44, to obtain the autoencoder long short-term memory parallel model;

[0054] Step S46: Model training, specifically, training the autoencoder long short-term memory parallel model to obtain an elevator fault prediction model;

[0055] Step S47: Calculate the elevator fault prediction result, specifically, perform elevator fault prediction through the elevator fault prediction model to obtain elevator fault prediction information.

[0056] Further, in step S5, the elevator safety engineering optimization is specifically to optimize the elevator operation parameters and maintenance plan by integrating the elevator operation status evaluation information and the elevator fault prediction information, so as to achieve the elevator safety engineering optimization;

[0057] Dynamically adjust elevator operating parameters based on elevator operating status assessment information and elevator fault prediction information, wherein the elevator operating parameters include elevator speed, acceleration, deceleration and load limit, so as to reduce the intensity of elevator operation under high-risk conditions, thereby reducing the probability of elevator failure;

[0058] Adjust the maintenance frequency of elevators based on the elevator operation status assessment information and elevator failure prediction information, and give priority to high-risk elevators.

[0059] The present invention provides an artificial intelligence-based safety engineering optimization system, comprising: a data preparation module, a data preprocessing module, an elevator operation status evaluation module, an elevator fault prediction module and an elevator safety engineering optimization module;

[0060] The data preparation module is used for data preparation, obtains elevator safety original data through data preparation, and sends the elevator safety original data to the data preprocessing module;

[0061] The data preprocessing module is used for data preprocessing, and obtains elevator status integrated data and elevator fault integrated data through data preprocessing, and sends the elevator status integrated data to the elevator operation status evaluation module, and sends the elevator fault integrated data to the elevator fault prediction module;

[0062] The elevator operation status evaluation module is used for evaluating the elevator operation status, obtaining elevator operation status evaluation information through the elevator operation status evaluation, and sending the elevator operation status evaluation information to the elevator fault prediction module and the elevator safety engineering optimization module;

[0063] The elevator fault prediction module is used for elevator fault prediction, obtains elevator fault prediction information through elevator fault prediction, and sends the elevator fault prediction information to the elevator safety engineering optimization module;

[0064] The elevator safety engineering optimization module integrates elevator operation status evaluation information and elevator fault prediction information, optimizes elevator operation parameters and maintenance plans, and realizes elevator safety engineering optimization.

[0065] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0066] (1) In view of the technical problem that the elevator operation status assessment has high requirements for model prediction accuracy in the existing safety engineering optimization process and needs further parameter optimization, and the traditional sparrow search algorithm has limited parameter optimization capability, resulting in poor model prediction effect and affecting the reliability of elevator operation status assessment, this scheme creatively adopts the long short-term memory model with enhanced parameter optimization to evaluate the elevator operation status. By introducing tent chaos mapping, adaptive inertia weight and random inertia weight, the diversity of sparrow population is increased, which effectively avoids falling into the local optimal solution, improves the global search capability of the algorithm, better optimizes parameters, improves the model prediction effect, and thus enhances the reliability of elevator operation status assessment;

[0067] (2) In view of the technical problems in the existing safety engineering optimization process, the elevator fault data is complex and often contains noise. The prediction accuracy and robustness of a single model in elevator fault prediction are poor, which easily leads to false alarms and increases unnecessary maintenance costs. This solution creatively adopts the autoencoder long short-term memory parallel model for elevator fault prediction. By processing the autoencoder and the long short-term memory network in parallel, it better captures rich feature representations and improves the model prediction accuracy. The gradient boosting technology is used for integrated learning to enhance the model robustness, thereby reducing the false alarm rate and maintenance costs.

[0068] (3) In view of the technical problems in the existing safety engineering optimization process, traditional elevator safety engineering optimization relies on the experience and judgment of professional technicians, is easily affected by subjective factors, and requires a lot of time and human resources. The process is time-consuming and labor-intensive, resulting in low efficiency and poor effectiveness of elevator safety engineering optimization. This solution creatively integrates elevator operation status evaluation information and elevator fault prediction information to achieve elevator safety engineering optimization, reduce the influence of subjective factors, effectively improve the overall stability and reliability of the elevator system, and promote the intelligent development of elevator system management. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 A schematic diagram of a safety engineering optimization method based on artificial intelligence provided by the present invention;

[0070] Figure 2 A schematic diagram of a safety engineering optimization system based on artificial intelligence provided by the present invention;

[0071] Figure 3 is a schematic flow chart of step S2;

[0072] Figure 4 is a schematic flow chart of step S3;

[0073] Figure 5 It is a schematic diagram of the process of step S4.

[0074] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0075] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0076] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0077] Example 1, see Figure 1The present invention provides a safety engineering optimization method based on artificial intelligence, which comprises the following steps:

[0078] Step S1: data preparation;

[0079] Step S2: data preprocessing;

[0080] Step S3: Elevator operation status evaluation;

[0081] Step S4: Elevator failure prediction;

[0082] Step S5: Elevator safety engineering optimization.

[0083] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, the data preparation is used to collect the original data required for elevator safety engineering optimization, specifically by collecting elevator system related data to obtain elevator safety original data, and the elevator safety original data includes elevator operation data, environmental data, historical fault data and maintenance records.

[0084] Example 3, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S2, the data preprocessing is used to preprocess the elevator safety original data. Specifically, the elevator state integrated data and the elevator fault integrated data are obtained by preprocessing the elevator safety original data, including the following steps:

[0085] Step S21: data cleaning, specifically, cleaning the original elevator safety data by deleting outliers, filling missing values, and removing duplicate values ​​to obtain elevator safety cleansing data;

[0086] Step S22: normalization processing, specifically, performing data normalization processing on the elevator safety cleaning data to obtain elevator safety standard data;

[0087] Step S23: Data integration, specifically, integrating the elevator operation data and environmental data in the elevator safety standard data to obtain elevator status integrated data, and integrating the historical fault data and maintenance records in the elevator safety standard data to obtain elevator fault integrated data.

[0088] Example 4, see Figure 1 and Figure 4 This embodiment is based on the above embodiment. In step S3, the elevator operation status evaluation is used to evaluate the current operation status of the elevator. Specifically, based on the elevator state integration data, the long short-term memory model with enhanced parameter optimization is used to evaluate the elevator operation status to obtain the elevator operation status evaluation information, including steps S31-S34:

[0089] Step S31: constructing a long short-term memory model, specifically constructing a long short-term memory model including an input gate, a forget gate, a memory unit and an output gate to obtain a long short-term memory model;

[0090] Step S32: improving the sparrow search algorithm, specifically, introducing tent chaos mapping, adaptive inertia weight and random inertia weight into the sparrow search algorithm, improving the sparrow search algorithm, and obtaining an improved sparrow search algorithm;

[0091] The improved sparrow search algorithm comprises the following steps:

[0092] Step S321: Initializing the sparrow population through tent chaos mapping to increase the diversity of initial solutions, including steps S3211-S3212:

[0093] Step S3211: Randomly generate each sparrow position in the interval [0,1], and the calculation formula is:

[0094] ;

[0095] In the formula, x ij is the position of the i-th sparrow in the j-th dimension, i is the sparrow index, j is the solution space dimension index, d is the matrix dimension, which is used to represent the solution space dimension, which is equal to the total number of model parameters that need to be optimized, rand(0,1) is a random number generated in the interval [0,1], pos i is the i-th sparrow position, x i1 is the position of the i-th sparrow in the first dimension, x i2 is the position of the i-th sparrow in the second dimension, x id is the position of the i-th sparrow in the d-th dimension;

[0096] Step S3212: Perform tent chaos mapping on each sparrow position until the maximum mapping times are reached to obtain the initial population of sparrows;

[0097] The calculation formula of the tent chaos map is:

[0098] ;

[0099] Where k is the mapping index, is the position of the i-th sparrow in the j-th dimension of the k+1-th mapping, is the position of the i-th sparrow in the k-th mapping on the j-th dimension;

[0100] Step S322: construct a fitness function to evaluate the quality of each sparrow position, specifically, use the cross entropy loss function as the fitness function, and calculate the fitness value of each sparrow position through the fitness function;

[0101] Step S323: updating the position of the finder, specifically introducing an adaptive inertia weight in the updating of the position of the finder, and updating the position of the finder. The finder specifically refers to a sparrow looking for food. The calculation formula for updating the position of the finder is:

[0102] ;

[0103] In the formula, is the adaptive inertia weight, exp(·) is the natural exponential function, t is the iteration index, and t max is the maximum number of iterations, is the position of the i-th sparrow at the t+1th iteration, is the position of the i-th sparrow at the t-th iteration, is an adjustment parameter with a value range of [0, 1], used to adjust the amplitude of position update, R1 is a warning value with a value range of [0, 1], val is a safety value with a value range of [0.5, 1], R2 is a random number that follows a normal distribution, L is a 1×d matrix with all elements being 1, d is the matrix dimension, used to represent the dimension of the solution space;

[0104] Step S324: Follower position update, specifically introducing random inertia weight in follower position update, and performing follower position update. The follower specifically refers to a sparrow that follows the finder to find food. The calculation formula for the follower position update is:

[0105] ;

[0106] In the formula, is the random inertia weight, is the random inertia weight mean, is the standard deviation of the random inertia weight, which is used to control the fluctuation range of the random inertia weight. N (0, 1) is a random number that obeys the standard normal distribution. is the best position occupied by the finder at the t+1th iteration, is the global worst position at the tth iteration, A is a 1×d matrix with values ​​of {-1, 1}, and T is the transpose operation;

[0107] Step S325: The position of the alerter is updated. The alerter specifically refers to the sparrow responsible for identifying potential dangers. The calculation formula is:

[0108] ;

[0109] In the formula, is the global optimal position at the tth iteration, f i is the fitness value of the i-th sparrow position, f gis the fitness value of the global optimal position, f w is the fitness value of the global worst position, is a very small constant used to avoid the situation where the denominator is zero, and R3 is a random number uniformly distributed in the interval [-1, 1];

[0110] Step S326: updating the position of the sparrow by repeatedly executing steps S323 to S325 until the maximum number of iterations is reached, and calculating the global optimal position as the optimal parameter of the model by iteration, wherein the maximum number of iterations is 500;

[0111] Step S33: constructing a long short-term memory model with enhanced parameter optimization, specifically by training the long short-term memory model and using an improved sparrow search algorithm to perform parameter tuning to obtain a long short-term memory model with enhanced parameter optimization;

[0112] Step S34: Calculate the elevator operation status evaluation result, specifically, based on the elevator state integration data, perform the elevator operation status evaluation through the long short-term memory model with enhanced parameter optimization to obtain the elevator operation status evaluation information;

[0113] By performing the above operations, in view of the technical problem that in the existing safety engineering optimization process, the elevator operation status assessment has high requirements for model prediction accuracy and needs further parameter optimization, while the parameter optimization capability of the traditional sparrow search algorithm is limited, resulting in poor model prediction effect and affecting the reliability of the elevator operation status assessment, this solution creatively adopts the long short-term memory model with enhanced parameter optimization to evaluate the elevator operation status. By introducing tent chaos mapping, adaptive inertia weight and random inertia weight, the diversity of the sparrow population is increased, effectively avoiding falling into the local optimal solution, improving the global search capability of the algorithm, better optimizing parameters, improving the model prediction effect, and thus enhancing the reliability of the elevator operation status assessment.

[0114] Example 5, see Figure 1 and Figure 5 , this embodiment is based on the above embodiment. In step S4, the elevator fault prediction is used to predict the probability of future elevator failures. Specifically, based on the elevator fault integration data and the elevator operation status evaluation information, the autoencoder long short-term memory parallel model is used to perform elevator fault prediction to obtain elevator fault prediction information;

[0115] The autoencoder long short-term memory parallel model includes an input layer, a parallel autoencoder long short-term memory layer, a gradient boosting prediction layer and an output layer;

[0116] The parallel autoencoder long short-term memory layer includes a parallel autoencoder subnet and a long short-term memory subnet;

[0117] The autoencoder subnet is used to learn data features in an unsupervised manner;

[0118] The long short-term memory subnetwork is used to process time series data and capture long-term dependencies;

[0119] The gradient boosting prediction layer is used to comprehensively process the output of the long short-term memory layer of the parallel autoencoder and generate a prediction result;

[0120] The elevator failure prediction comprises the following steps:

[0121] Step S41: constructing an input layer for receiving input data, specifically receiving elevator fault integration data and elevator operation status evaluation information as input data of the model through the input layer;

[0122] Step S42: Constructing a parallel autoencoder long short-term memory layer, including steps S421-S422:

[0123] Step S421: construct an autoencoder subnet, specifically construct an autoencoder subnet including an encoder, a repeated vector layer, a decoder, and a reconstruction loss layer, and calculate the input data through the autoencoder subnet to obtain the output features of the autoencoder subnet;

[0124] Step S422: constructing a long short-term memory subnet, specifically, using a long short-term memory model including an input gate, a forget gate, a memory unit, and an output gate as a long short-term memory subnet, calculating the input data through the long short-term memory subnet, and obtaining the long short-term memory subnet output feature;

[0125] The calculation formula of the long short-term memory subnet is:

[0126] ;

[0127] Where v is the time step index, which is used to represent the current moment, ig v is the output vector of the input gate at the current moment, Sig(·) is the S-type activation function, and W ig is the input gate weight, h v-1 is the hidden state at the previous moment, x v is the input data at the current moment, b ig is the input gate bias term, fg v is the output vector of the forget gate at the current moment, W fg is the forget gate weight, b fg is the forget gate bias term, C v is the current state of the memory unit, C v-1 is the state of the memory unit at the previous moment, is the candidate memory cell state, og v is the output vector of the output gate at the current moment, Wog is the output gate weight, b og is the output gate bias term;

[0128] Step S43: constructing a gradient boosting prediction layer, specifically using a lightweight gradient boosting model as the gradient boosting prediction layer, and integrating the output features of the autoencoder subnet and the output features of the long short-term memory subnet through the gradient boosting prediction layer to obtain the final prediction result;

[0129] Step S44: construct an output layer for outputting the final prediction result;

[0130] Step S45: constructing an autoencoder long short-term memory parallel model, specifically, constructing the autoencoder long short-term memory parallel model through the input layer of step S41, the parallel autoencoder long short-term memory layer of step S42, the gradient boosting prediction layer of step S43 and the output layer of step S44, to obtain the autoencoder long short-term memory parallel model;

[0131] Step S46: Model training, specifically, training the autoencoder long short-term memory parallel model to obtain an elevator fault prediction model;

[0132] Step S47: calculating the elevator fault prediction result, specifically performing elevator fault prediction through the elevator fault prediction model to obtain elevator fault prediction information;

[0133] By performing the above operations, in order to solve the technical problem that in the existing safety engineering optimization process, elevator fault data is complex and often contains noise, and the prediction accuracy and robustness of a single model in elevator fault prediction are poor, which is easy to cause false alarms and increase unnecessary maintenance costs, this solution creatively adopts the autoencoder long short-term memory parallel model to predict elevator faults. By processing the autoencoder and the long short-term memory network in parallel, it can better capture rich feature representations and improve the model prediction accuracy. Through integrated learning through gradient boosting technology, the model robustness is enhanced, thereby reducing the false alarm rate and maintenance costs.

[0134] Example 6, see Figure 1 , this embodiment is based on the above embodiment. In step S5, the elevator safety engineering optimization is specifically to optimize the elevator operation parameters and maintenance plan by integrating the elevator operation status evaluation information and the elevator fault prediction information, so as to realize the elevator safety engineering optimization;

[0135] Dynamically adjust elevator operating parameters based on elevator operating status assessment information and elevator fault prediction information, wherein the elevator operating parameters include elevator speed, acceleration, deceleration and load limit, so as to reduce the intensity of elevator operation under high-risk conditions, thereby reducing the probability of elevator failure;

[0136] Adjust the maintenance frequency of elevators based on the elevator operation status assessment information and elevator failure prediction information, and give priority to high-risk elevators.

[0137] By performing the above operations, in order to address the technical problems in the existing safety engineering optimization process, such as traditional elevator safety engineering optimization relies on the experience and judgment of professional technicians, is easily affected by subjective factors, requires a lot of time and human resources, and is time-consuming and labor-intensive, resulting in low efficiency and poor effectiveness of elevator safety engineering optimization, this solution creatively integrates elevator operation status evaluation information and elevator fault prediction information to achieve elevator safety engineering optimization, reduce the influence of subjective factors, effectively improve the overall stability and reliability of the elevator system, and promote the intelligent development of elevator system management.

[0138] Embodiment 7, see Figure 2 , this embodiment is based on the above embodiment, and the present invention provides an artificial intelligence-based safety engineering optimization system, including: a data preparation module, a data preprocessing module, an elevator operation status evaluation module, an elevator fault prediction module and an elevator safety engineering optimization module;

[0139] The data preparation module is used for data preparation, obtains elevator safety original data through data preparation, and sends the elevator safety original data to the data preprocessing module;

[0140] The data preprocessing module is used for data preprocessing, and obtains elevator status integrated data and elevator fault integrated data through data preprocessing, and sends the elevator status integrated data to the elevator operation status evaluation module, and sends the elevator fault integrated data to the elevator fault prediction module;

[0141] The elevator operation status evaluation module is used for evaluating the elevator operation status, obtaining elevator operation status evaluation information through the elevator operation status evaluation, and sending the elevator operation status evaluation information to the elevator fault prediction module and the elevator safety engineering optimization module;

[0142] The elevator fault prediction module is used for elevator fault prediction, obtains elevator fault prediction information through elevator fault prediction, and sends the elevator fault prediction information to the elevator safety engineering optimization module;

[0143] The elevator safety engineering optimization module integrates elevator operation status evaluation information and elevator fault prediction information, optimizes elevator operation parameters and maintenance plans, and realizes elevator safety engineering optimization.

[0144] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0145] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0146] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A safety engineering optimization method based on artificial intelligence, characterized in that: The method comprises the following steps: Step S1: data preparation, used to collect the original data required for elevator safety engineering optimization and obtain elevator safety original data; Step S2: data preprocessing, which is used to preprocess the original elevator safety data, specifically, obtaining elevator status integrated data and elevator fault integrated data by performing data preprocessing on the original elevator safety data; Step S3: Elevator operation status evaluation, used to evaluate the current operation status of the elevator, specifically, based on the elevator state integration data, using the long short-term memory model with enhanced parameter optimization to evaluate the elevator operation status, and obtain the elevator operation status evaluation information, including steps S31-S34: Step S31: constructing a long short-term memory model, specifically constructing a long short-term memory model including an input gate, a forget gate, a memory unit and an output gate to obtain a long short-term memory model; Step S32: improving the sparrow search algorithm, specifically, introducing tent chaos mapping, adaptive inertia weight and random inertia weight into the sparrow search algorithm, improving the sparrow search algorithm, and obtaining an improved sparrow search algorithm; The improved sparrow search algorithm comprises the following steps: Step S321: Initializing the sparrow population through tent chaos mapping, including steps S3211-S3212: Step S3211: Randomly generate each sparrow position in the interval [0,1], and the calculation formula is: ； In the formula, x ij is the position of the i-th sparrow in the j-th dimension, i is the sparrow index, j is the solution space dimension index, d is the matrix dimension, which is used to represent the solution space dimension, which is equal to the total number of model parameters that need to be optimized, rand(0,1) is a random number generated in the interval [0,1], pos i is the i-th sparrow position, x i1 is the position of the i-th sparrow in the first dimension, x i2 is the position of the i-th sparrow in the second dimension, x id is the position of the i-th sparrow in the d-th dimension; Step S3212: Perform tent chaos mapping on each sparrow position until the maximum mapping times are reached to obtain the initial population of sparrows; The calculation formula of the tent chaos map is: ； Where k is the mapping index, is the position of the i-th sparrow in the j-th dimension of the k+1-th mapping, is the position of the i-th sparrow in the k-th mapping on the j-th dimension; Step S322: construct a fitness function to evaluate the quality of each sparrow position, specifically, use the cross entropy loss function as the fitness function, and calculate the fitness value of each sparrow position through the fitness function; Step S323: updating the position of the finder, specifically introducing an adaptive inertia weight in the updating of the position of the finder, and updating the position of the finder. The finder specifically refers to a sparrow looking for food. The calculation formula for updating the position of the finder is: ； In the formula, is the adaptive inertia weight, exp(·) is the natural exponential function, t is the iteration index, and t max is the maximum number of iterations, is the position of the i-th sparrow at the t+1th iteration, is the position of the i-th sparrow at the t-th iteration, is an adjustment parameter with a value range of [0, 1], used to adjust the amplitude of position update, R1 is a warning value with a value range of [0, 1], val is a safety value with a value range of [0.5, 1], R2 is a random number that follows a normal distribution, L is a 1×d matrix with all elements being 1, d is the matrix dimension, used to represent the dimension of the solution space; Step S324: Follower position update, specifically introducing random inertia weight in follower position update, and performing follower position update. The follower specifically refers to a sparrow that follows the finder to find food. The calculation formula for the follower position update is: ； In the formula, is the random inertia weight, is the random inertia weight mean, is the standard deviation of the random inertia weight, which is used to control the fluctuation range of the random inertia weight. N (0, 1) is a random number that obeys the standard normal distribution. is the best position occupied by the finder at the t+1th iteration, is the global worst position at the tth iteration, A is a 1×d matrix with values ​​of {-1, 1}, and T is the transpose operation; Step S325: The position of the alerter is updated. The alerter specifically refers to the sparrow responsible for identifying potential dangers. The calculation formula is: ； In the formula, is the global optimal position at the tth iteration, f i is the fitness value of the i-th sparrow position, f g is the fitness value of the global optimal position, f w is the fitness value of the global worst position, is a very small constant used to avoid the situation where the denominator is zero, and R3 is a random number uniformly distributed in the interval [-1, 1]; Step S326: updating the position of the sparrow by repeatedly executing steps S323 to S325 until the maximum number of iterations is reached, and calculating the global optimal position as the optimal parameter of the model by iteration, wherein the maximum number of iterations is 500; Step S33: constructing a long short-term memory model with enhanced parameter optimization, specifically by training the long short-term memory model and using an improved sparrow search algorithm to perform parameter tuning to obtain a long short-term memory model with enhanced parameter optimization; Step S34: Calculate the elevator operation status evaluation result, specifically, based on the elevator state integration data, perform the elevator operation status evaluation through the long short-term memory model with enhanced parameter optimization to obtain the elevator operation status evaluation information; Step S4: Elevator fault prediction, which is used to predict the probability of future elevator failures. Specifically, based on the elevator fault integration data and the elevator operation status evaluation information, an autoencoder long short-term memory parallel model is used to perform elevator fault prediction to obtain elevator fault prediction information; The autoencoder long short-term memory parallel model includes an input layer, a parallel autoencoder long short-term memory layer, a gradient boosting prediction layer and an output layer; the parallel autoencoder long short-term memory layer includes a parallel autoencoder subnet and a long short-term memory subnet; the autoencoder subnet is used to learn data features in an unsupervised manner; the long short-term memory subnet is used to process time series data and capture long-term dependencies; the gradient boosting prediction layer is used to comprehensively process the output of the parallel autoencoder long short-term memory layer and generate prediction results; The elevator failure prediction comprises the following steps: Step S41: constructing an input layer for receiving input data, specifically receiving elevator fault integration data and elevator operation status evaluation information as input data of the model through the input layer; Step S42: Constructing a parallel autoencoder long short-term memory layer, including steps S421-S422: Step S421: construct an autoencoder subnet, specifically construct an autoencoder subnet including an encoder, a repeated vector layer, a decoder, and a reconstruction loss layer, and calculate the input data through the autoencoder subnet to obtain the output features of the autoencoder subnet; Step S422: constructing a long short-term memory subnet, specifically, using a long short-term memory model including an input gate, a forget gate, a memory unit, and an output gate as a long short-term memory subnet, calculating the input data through the long short-term memory subnet, and obtaining the long short-term memory subnet output feature; Step S43: constructing a gradient boosting prediction layer, specifically using a lightweight gradient boosting model as the gradient boosting prediction layer, and integrating the output features of the autoencoder subnet and the output features of the long short-term memory subnet through the gradient boosting prediction layer to obtain the final prediction result; Step S44: construct an output layer for outputting the final prediction result; Step S45: constructing an autoencoder long short-term memory parallel model, specifically, constructing the autoencoder long short-term memory parallel model through the input layer of step S41, the parallel autoencoder long short-term memory layer of step S42, the gradient boosting prediction layer of step S43 and the output layer of step S44, to obtain the autoencoder long short-term memory parallel model; Step S46: Model training, specifically, training the autoencoder long short-term memory parallel model to obtain an elevator fault prediction model; Step S47: calculating the elevator fault prediction result, specifically performing elevator fault prediction through the elevator fault prediction model to obtain elevator fault prediction information; Step S5: Elevator safety engineering optimization, specifically, integrating elevator operation status evaluation information and elevator fault prediction information, optimizing elevator operation parameters and maintenance plans, and realizing elevator safety engineering optimization.

2. The artificial intelligence-based safety engineering optimization method according to claim 1, characterized in that: The elevator safety raw data includes elevator operation data, environmental data, historical fault data and maintenance records.

3. The artificial intelligence-based safety engineering optimization method according to claim 2 is characterized in that: In step S2, the data preprocessing is used to preprocess the elevator safety raw data, including the following steps: Step S21: data cleaning, specifically, cleaning the original elevator safety data by deleting outliers, filling missing values, and removing duplicate values ​​to obtain elevator safety cleansing data; Step S22: normalization processing, specifically, performing data normalization processing on the elevator safety cleaning data to obtain elevator safety standard data; Step S23: Data integration, specifically, integrating the elevator operation data and environmental data in the elevator safety standard data to obtain elevator status integrated data, and integrating the historical fault data and maintenance records in the elevator safety standard data to obtain elevator fault integrated data.

4. A safety engineering optimization system based on artificial intelligence, used to implement a safety engineering optimization method based on artificial intelligence as described in any one of claims 1 to 3, characterized in that: It includes data preparation module, data preprocessing module, elevator operation status assessment module, elevator fault prediction module and elevator safety engineering optimization module; The data preparation module is used for data preparation, obtains elevator safety original data through data preparation, and sends the elevator safety original data to the data preprocessing module; The data preprocessing module is used for data preprocessing, and obtains elevator status integrated data and elevator fault integrated data through data preprocessing, and sends the elevator status integrated data to the elevator operation status evaluation module, and sends the elevator fault integrated data to the elevator fault prediction module; The elevator operation status evaluation module is used for evaluating the elevator operation status, obtaining elevator operation status evaluation information through the elevator operation status evaluation, and sending the elevator operation status evaluation information to the elevator fault prediction module and the elevator safety engineering optimization module; The elevator fault prediction module is used for elevator fault prediction, obtains elevator fault prediction information through elevator fault prediction, and sends the elevator fault prediction information to the elevator safety engineering optimization module; The elevator safety engineering optimization module integrates elevator operation status evaluation information and elevator fault prediction information, optimizes elevator operation parameters and maintenance plans, and realizes elevator safety engineering optimization.

Images