Process pipeline installation quality evaluation method and device

Abstract

The invention relates to a process pipeline installation quality evaluation method and device, and the method comprises the steps: collecting first-hand data of a construction site, and screening the first-hand data to obtain evaluation index data related to the quality of a process pipeline installation project; based on a pre-established process pipeline installation project quality evaluation index system table, screening from the evaluation index data to obtain a plurality of target second-level indexes under each first-level index; and obtaining evaluation result data of each target second-level index under each first-level index, and inputting each evaluation result data under each first-level index into the process pipeline installation quality evaluation model to obtain an installation quality scoring result. According to the invention, quantitative evaluation of the installation quality of the process pipeline can be realized, and the method has a certain guiding effect on quality control and management of the installation project of the process pipeline.

Description

Technical Field

[0001] The present invention relates to a method and device for evaluating the installation quality of process pipelines. Background Art

[0002] Process pipelines are important equipment indispensable in the engineering construction of petrochemical plants, and their installation quality directly affects production efficiency and product quality. Therefore, it is very important to evaluate the installation quality of process pipelines. Summary of the Invention

[0003] In order to quantitatively evaluate the installation quality status of process pipelines and achieve quality control and management of process pipeline installation projects, the present invention proposes a method and device for evaluating the installation quality of process pipelines. The technical solutions proposed by the present invention are as follows:

[0004] In a first aspect, the present invention provides a method for evaluating the installation quality of process pipelines, including:

[0005] Collect first-hand data at the construction site, and screen out evaluation index data related to the installation quality of process pipeline projects from the first-hand data;

[0006] Based on a pre-established quality evaluation index system table for process pipeline installation projects, screen out multiple target secondary indicators under each primary indicator from the evaluation index data;

[0007] Obtain evaluation result data of each of the target secondary indicators under each primary indicator, and input the evaluation result data under each primary indicator into a process pipeline installation quality evaluation model to obtain an installation quality score result.

[0008] In one or some embodiments, the process pipeline installation quality evaluation model includes a post-evaluation model and evaluation sub-models corresponding to each primary indicator;

[0009] The step of inputting the evaluation result data under each primary indicator into the process pipeline installation quality evaluation model to obtain an installation quality score result includes:

[0010] For each primary indicator, input the evaluation result data under the primary indicator into the evaluation sub-model corresponding to the primary indicator to obtain a corresponding primary evaluation result;

[0011] Input the primary evaluation results corresponding to each primary indicator into the post-evaluation model to obtain the installation quality score result.

[0012] In one or some embodiments, the evaluation sub-model is trained by the following method:

[0013] Obtain a quality evaluation index database for process pipeline installation projects;

[0014] For each first-level index in the quality evaluation index database of the process pipeline installation project, a corresponding BP neural network is established respectively; wherein, the number of neurons in the input layer of the BP neural network is the number of second-level indexes under the corresponding first-level index;

[0015] The quality evaluation index database of the process pipeline installation project is divided according to the first-level indexes to obtain the data set corresponding to each first-level index;

[0016] For each first-level index, based on the data set corresponding to the first-level index, the BP neural network corresponding to the first-level index is trained using the grid search and cross-validation methods to determine the optimal parameters, and the evaluation sub-model is obtained.

[0017] In one or some embodiments, the step of, for each first-level index, based on the data set corresponding to the first-level index, training the BP neural network corresponding to the first-level index using the grid search and cross-validation methods to determine the optimal parameters, and obtaining the evaluation sub-model includes:

[0018] For each first-level index, the corresponding BP neural network is trained using the data set corresponding to the first-level index to determine the optimal number of nodes; wherein, the optimal number of nodes is the number of neurons in the hidden layer of the BP neural network;

[0019] The BP neural network is updated based on the optimal number of nodes to obtain an improved network model;

[0020] The improved network model is iteratively trained based on the data set corresponding to the first-level index, the grid search is used for parameter search, and the cross-validation is used to evaluate the performance of the model until the performance of the model meets the preset requirements, and the evaluation sub-model is obtained.

[0021] In one or some embodiments, the post-evaluation model is trained in the following manner:

[0022] After training the evaluation sub-models corresponding to each first-level index, the quality evaluation index database of the process pipeline installation project is divided into a training set and a test set;

[0023] According to the first-level indexes, the training set is input into the corresponding evaluation sub-models to obtain the corresponding first-level evaluation result sets;

[0024] The initial post-evaluation model is trained using the first-level evaluation result sets, and the weights of each neuron in the input layer of the initial post-evaluation model are adjusted using the gradient descent method to obtain the trained model;

[0025] Use the test set to perform cross - validation on the trained model, repeat the process of adjusting the weights until the preset conditions are met, and obtain the post - evaluation model.

[0026] In one or some embodiments, inputting the first - level evaluation results corresponding to each first - level index into the post - evaluation model to obtain the installation quality scoring result includes:

[0027] Input the first - level evaluation results corresponding to each first - level index into the post - evaluation model as follows to obtain the installation quality scoring result:

[0028] I = λ 1 X 1 +...+λ i X i +...+λ n X n

[0029] In the formula, I is the installation quality scoring result, λ 1 ,...,λ i ,...,λ n are the weight coefficients of each first - level index respectively, and X 1 ,...,X i ,...,X n are the first - level evaluation results of each first - level index respectively.

[0030] In one or some embodiments, collecting the first - hand data at the construction site and screening out the evaluation index data related to the quality of process pipeline installation projects includes:

[0031] Collect the first - hand data at the construction site, and screen out all the secondary index data related to the installation data quality control index, installation material quality control index, installation prefabrication quality control index, installation construction quality control index, and installation inspection quality control index from the first - hand data to obtain the evaluation index data.

[0032] In one or some embodiments, based on the pre - established quality evaluation index system table for process pipeline installation projects, screening out multiple target secondary indicators under each first - level index from the evaluation index data includes:

[0033] For the installation data quality control index, collect the design disclosure, drawing review, equipment unpacking inspection record, delivery data, final acceptance, and construction plan submission data from the evaluation index data;

[0034] For the quality control indicators of installation materials, collect the data of on-site assessment of welders, registration of qualified welders, submission of qualified welders and welding procedure qualifications, compliance inspection of welding procedure qualifications, and submission of engineering materials / equipment / fittings from the evaluation index data;

[0035] For the quality control indicators of installation prefabrication, collect the data of process handover, heat treatment of prefabricated welds, hardness test of prefabricated pipes, installation welding of prefabricated pipes, and appearance inspection of welds from the evaluation index data;

[0036] For the quality control indicators of installation construction, collect the data of heat treatment of pipeline installation welds, hardness of pipeline installation, non-destructive testing of pipeline installation, repair and re-inspection of pipeline installation welds, inspection of installation of pipeline compensation devices, installation of fixed / sliding pipe supports, adjustment of spring supports and hangers, installation of safety accessories, commissioning records of safety valves, welding records of pressure pipelines, static grounding test of pipelines, and installation data of pipelines connecting to machines from the evaluation index data;

[0037] For the quality control indicators of installation inspection, collect the data of inspection of contact lines of metal ring gaskets / lens gaskets, closed water test of drainage pipelines, notification and supervision and inspection of pressure pipeline installation, technical disclosure, inspection records of welding consumables warehousing, temperature / humidity records of welding consumables warehouse, electrode drying records, electrode distribution and recovery records, inspection of materials entering the site, and on-site meteorological records from the evaluation index data.

[0038] In one or some embodiments, the method further includes:

[0039] According to the pre-established quality specifications, conduct a grade evaluation on the installation quality scoring results to obtain a quality rating result.

[0040] In a second aspect, the present invention provides a device for evaluating the installation quality of process pipelines, including:

[0041] A collection module, configured to collect first-hand data at the construction site and screen out evaluation index data related to the installation quality of process pipeline projects from the first-hand data;

[0042] A screening module, configured to screen out multiple target secondary indicators under each first-level indicator from the evaluation index data based on the pre-established evaluation index system table for the installation quality of process pipeline projects;

[0043] An evaluation module, configured to obtain the evaluation result data of each of the target secondary indicators under each first-level indicator, input the evaluation result data under each first-level indicator into the installation quality evaluation model of process pipelines, and obtain an installation quality scoring result.

[0044] In a third aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the method for evaluating the installation quality of a process pipeline as described in the first aspect is implemented.

[0045] In a fourth aspect, the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus;

[0046] The memory is used to store a computer program;

[0047] The processor is used to implement the method for evaluating the installation quality of a process pipeline as described in the first aspect when executing the program stored on the memory.

[0048] Based on the above technical solutions, the beneficial effects of the present invention compared with the prior art are as follows:

[0049] The method for evaluating the installation quality of a process pipeline provided by the present invention collects first-hand data at the construction site, and screens out evaluation index data related to the installation quality of the process pipeline from the first-hand data; secondly, based on the pre-established evaluation index system table for the installation quality of the process pipeline, multiple target secondary indicators under each first-level indicator are screened out from the evaluation index data; then, the evaluation result data of the target secondary indicators is obtained, and the evaluation result data corresponding to each first-level indicator is evaluated based on the pre-established installation quality evaluation model of the process pipeline to obtain the installation quality score result. By scoring the quality status of the process pipeline installation project, the quality situation of the process pipeline installation project is reflected, which has a certain guiding role for the quality control and management of the process pipeline installation project.

[0050] Other features and advantages of the present invention will be described in the following specification, and part of them will become obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention are realized and obtained by the structures specifically pointed out in the specification, the claims, and the drawings.

[0051] To make the above objectives, features, and advantages of the present invention more obvious and understandable, the following specifically enumerates preferred embodiments and, in conjunction with the accompanying drawings, makes a detailed description as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts;

[0053] Figure 1 It is a schematic flowchart of the process pipeline installation quality evaluation method provided by an embodiment of the present invention;

[0054] Figure 2 It is a structural framework diagram of the BP neural network provided by an embodiment of the present invention;

[0055] Figure 3 It is a schematic diagram of the input and output of the neurons of the BP neural network provided by an embodiment of the present invention;

[0056] Figure 4 It is an error performance curve diagram of the first-level indicators of the process pipeline installation project provided by an embodiment of the present invention;

[0057] Figure 5 It is a structural schematic diagram of the process pipeline installation quality evaluation device provided by an embodiment of the present invention;

[0058] Figure 6 It is a structural schematic diagram of the electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0059] Hereinafter, the exemplary embodiments of the present disclosure will be described in more detail with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be completely conveyed to those skilled in the art.

[0060] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present invention. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0061] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is 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 orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In addition, the terms "first", "second", and "third" are only used for descriptive purposes and should not be construed as indicating or implying relative importance.

[0062] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the terms "installation", "connection", and "coupling" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be a direct connection or an indirect connection through an intermediate medium, and it may be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0063] Process pipelines are important and indispensable equipment in the engineering construction of petrochemical plants. Their installation quality directly affects production efficiency and product quality. Therefore, it is very important to evaluate the installation quality of process pipelines. The inventor found in work that traditional installation quality evaluation methods mainly rely on manual experience and detection instruments, and there are problems such as low evaluation efficiency and low accuracy. With the development of computer technology, using neural networks to evaluate the installation quality of process pipelines has become a new method. In order to make the evaluation of the installation quality of process pipelines more accurate and applicable, the inventor conducted further research and studied the method for evaluating the installation quality of process pipelines based on the BP neural network, and made the present invention.

[0064] Embodiment 1

[0065] The present invention belongs to the field of engineering construction of petrochemical plants, and specifically relates to a method for evaluating the installation quality of process pipelines. Evaluating the installation quality status of process pipelines is an important part of the entire petrochemical engineering construction project. However, due to its own characteristics, it is not very easy to control the quality problems in the installation of process pipelines. This application proposes a method for evaluating the installation quality status of process pipelines based on the BP neural network. By calculating and obtaining the weight coefficients at all levels through model training, and combining the scores of each quality index by engineering experts, the final scoring method is obtained, and the installation quality status of the process pipeline is evaluated. The method for evaluating the installation quality of process pipelines provided in the embodiment of the present invention refers to Figure 1 as shown, including:

[0066] S101. Collect first-hand data at the construction site, and screen out evaluation index data related to the installation quality of the process pipeline project from the first-hand data;

[0067] Collect first-hand data at the construction site by means of digital handover. When collecting first-hand data, information on factors such as pipeline material quality, pipeline installation quality, operation and maintenance, environment, and design can be obtained through on-site investigations, inspection records, and actual measurements, so as to determine the key factors affecting the quality of the process pipeline. Screen out the main factors affecting the quality problems in the installation of the process pipeline from the first-hand data to obtain the evaluation index of the installation quality of the process pipeline project, that is, the above-mentioned evaluation index data.

[0068] S102. Based on the pre-established quality evaluation index system table for the process pipeline installation project, multiple target secondary indicators under each first-level indicator are screened from the evaluation index data;

[0069] Since the data indicators of review and observation are somewhat subjective, it is required that the experts of this project directly participate in writing the inspection and testing content and dividing the control point levels for different-level indicators to ensure the accuracy and effectiveness of the data. According to the characteristics of the process pipeline installation project and the quality control requirements in the actual construction process, combined with relevant industry standards (such as GB 50235-2010 "Code for Construction of Industrial Metal Pipelines" and GB 50184-2011 "Code for Acceptance of Construction Quality of Industrial Metal Pipelines", etc.), the quality evaluation index system table for the process pipeline installation project as shown in the following table is generated;

[0070]

[0071]

[0072] Table 1

[0073] Since among the secondary indicators corresponding to each first-level indicator in the evaluation index data, there are secondary indicators that do not belong to those in Table 1 above, it is necessary to screen the evaluation index data. For each first-level indicator, the evaluation index data is screened according to the secondary indicators in Table 1 to obtain the target secondary indicators.

[0074] S103. Obtain the evaluation result data of each of the target secondary indicators under each first-level indicator, and input the evaluation result data under each first-level indicator into the process pipeline installation quality evaluation model to obtain the installation quality score result.

[0075] The experts pre-evaluate each target secondary indicator and store the evaluation result data in the database. Taking the secondary indicator - design disclosure of the first-level indicator - quality control of process pipeline installation materials as an example, the experts pre-evaluate the design disclosure according to the design documents and requirements, and the corresponding evaluation result data. When conducting the process pipeline installation quality evaluation, from the evaluation result data of each target secondary indicator in the database, the evaluation result data under each first-level indicator is input into the process pipeline installation quality evaluation model to obtain the installation quality score result, realizing the quantitative evaluation of the process pipeline installation quality.

[0076] In an optional embodiment, the process pipeline installation quality evaluation model includes a post-evaluation model and evaluation sub-models corresponding to each first-level indicator; the output layer of the evaluation sub-models corresponding to each first-level indicator is connected to the input layer of the post-evaluation model, and the number of neurons in the input layer of the post-evaluation model is the number of evaluation sub-models.

[0077] Inputting the evaluation result data under each of the above-mentioned first-level indicators into the process pipeline installation quality evaluation model to obtain the installation quality score result, including:

[0078] S1031. For each first-level indicator, input the evaluation result data under the first-level indicator into the evaluation sub-model corresponding to the first-level indicator to obtain the corresponding first-level evaluation result;

[0079] Input the evaluation result data of each target second-level indicator under the first-level indicator into the evaluation sub-model corresponding to the first-level indicator. Among them, the evaluation result data of different target second-level indicators under the same first-level indicator are respectively input into different neurons of the input layer of the evaluation sub-model, and the sub-evaluation model evaluates all the evaluation result data of the first-level indicator to obtain the first-level evaluation result corresponding to the first-level indicator.

[0080] S1032. Input the first-level evaluation results corresponding to each first-level indicator into the post-evaluation model to obtain the installation quality score result.

[0081] Specifically, input the first-level evaluation results corresponding to each first-level indicator into the post-evaluation model as follows, perform weighted summation on the first-level evaluation results of each first-level indicator to obtain the installation quality score result:

[0082] I = λ 1 X 1 +... + λ i X i +... + λ n X n

[0083] In the formula, I is the installation quality score result, λ 1 ,..., λ i ,..., λ n are the weight coefficients of each first-level indicator respectively, and X 1 ,..., X i ,..., X n are the first-level evaluation results of each first-level indicator respectively.

[0084] In an optional embodiment, the method further includes:

[0085] S104. According to the pre-established quality specification, perform grade evaluation on the installation quality score result to obtain the quality rating result.

[0086] According to the provisions of the standard specifications and the actual requirements of the project, it can be known that: when 85 < I ≤ 100, the quality evaluation is excellent; when 70 < I ≤ 85, the quality evaluation is good; when I = 70, the quality evaluation is qualified; when I < 70, the quality evaluation is unqualified.

[0087] The present invention proposes a method for evaluating the installation quality of process pipelines based on a BP neural network. In an optional embodiment, the evaluation sub-model is trained in the following manner:

[0088] S201. Obtain the quality evaluation index database for the installation project of process pipelines;

[0089] Based on the quality evaluation index system table for the installation project of process pipelines as shown in Table 1 above, historical secondary index data is collected, and experts evaluate the historical secondary index data to obtain the corresponding evaluation result data, which is used as the quality evaluation index database for the installation project of process pipelines.

[0090] S202. For each first-level index in the quality evaluation index database for the installation project of process pipelines, establish a corresponding BP neural network respectively; wherein, the number of neurons in the input layer of the BP neural network is the number of secondary indexes under the corresponding first-level index;

[0091] For each first-level indicator, a corresponding BP neural network is established. When establishing the BP neural network, the number of neurons in the input layer of the BP neural network is set according to the number of second-level indicators under this first-level indicator in Table 1 above, and the number of neurons in the output layer is 1. Input layer: the input end of information, reading in the input data; Hidden layer: the information processing end, where the number of hidden layers can be set; Output layer: the output end of information, the required result. A three-layer improved BP (Back-Propagation) neural network is formed. The activation function of the hidden layer is the relu type function, and the activation function of the output layer is a linear function. The network training method uses the L-M (nonlinear least squares method) algorithm, and the root mean square error is used to measure the error degree between the neural network output value and the true value. Network training operation process: First, generate training samples, test samples, and validation samples according to the process pipeline installation engineering quality evaluation index database; Second, set parameters: including the number of neurons in the input layer, hidden layer, and output layer, activation function, initial weight value, initial learning rate, error rate, and expected target error, etc., determine the main influencing factors of the quality problems in the process pipeline installation, and based on the main influencing factors of the engineering quality problems in the process pipeline installation, determine the number of neurons in the input layer and output layer; Finally, start training the network. After the model test results are compared with the real data, if the error value is large, adjust the weights of the neurons through the error, further optimize the model parameters and then iterate the training again until it reaches within the range allowed by the expected target error, and the model training ends to obtain the weights corresponding to the model. In the present invention, grid search and cross-validation are used to obtain the number of hidden layers and the initial learning rate.

[0092] The present invention optimizes the existing BP neural network, and the optimized BP neural network refers to Figure 3 as shown: Each neuron is responsible for: input, judgment, and output. The neurons within each layer are not connected to each other, while the neurons between adjacent layers are connected to each other. Each neuron receives input information (matrix Xi) from other neurons, and each piece of information is transmitted through a neural wire with a weight (Wi). The neuron adds up these information to obtain a total input information (matrix [Xi+…+Xi+1]), compares the total input information with the threshold of the neuron, and then processes and judges through an "activation function" to obtain the final output information. This output information will be passed on layer by layer as the input information of the subsequent neurons until the last neuron outputs the expected information.

[0093] Therefore, the result of the j-th neuron in the l-th layer is as follows, where the activation function is represented by the symbol σ:

[0094]

[0095] Symbol convention: Denote the connection weights of k output neurons from the upper layer [l-1] of the network. b [l] is used to represent the bias of the neurons in this layer l, and is used to represent the linear result of the neurons in this layer l, and is used to represent the activation function outputs of k neurons in the upper layer [l-1]. Among them, define w [l] to represent the weight matrix, and each of its elements represents a weight, that is, each row is the weight connecting to the l-th layer:

[0096]

[0097] Similarly,

[0098]

[0099]

[0100] That is, the forward propagation process represents

[0101] The purpose of introducing the activation function is to introduce non-linearity into the model. If there is no activation function, that is, equivalent to the excitation function F(x)=x, then no matter how many layers your neural network has, it is ultimately a linear mapping, and then the approximation ability of the network is quite limited. A simple linear mapping cannot solve the problem of linear inseparability. Therefore, the present invention decides to introduce a non-linear function as the excitation function. The RELU function is a general activation function, which is improved for the shortcomings of the Sigmoid function and the Tanh function.

[0102] In the iterative training, during the iterative training process, the optimization of the model parameters is completed through error backpropagation. Refer to Figure 2 as shown, the specific steps are as follows:

[0103] Step 1: In each iteration process, input the sample data into the input layer of the neural network and process the information through the hidden layer.

[0104] Step 2: In the hidden layer, use the relu-type function as the activation function to perform a non-linear transformation on the input signal.

[0105] Step 3: Transmit the output of the hidden layer to the output layer and activate it through a linear function to obtain the final output value of the neural network.

[0106] Step 4: Compare the output value of the neural network with the true value, and calculate the root mean square error (RMSE) or other error metrics to evaluate the performance of the model.

[0107] Step 5: If the error exceeds the expected target error, adjust the weights of the neurons according to the error, that is, perform parameter optimization.

[0108] Step 6: The specific method of parameter optimization is to use the gradient descent method to update the network weights by calculating the error backpropagation, so that the error of the next iteration is smaller.

[0109] Step 7: Repeat the above steps until the error is within the allowable range of the expected target error or the preset maximum number of iterations is reached. The model training ends, and the weights corresponding to the model are obtained.

[0110] S203. Divide the quality evaluation index database of the process pipeline installation project according to the first-level indicators to obtain the data set corresponding to each first-level indicator.

[0111] S204. For each first-level indicator, based on the data set corresponding to the first-level indicator, use the grid search and cross-validation methods to train the BP neural network corresponding to the first-level indicator, determine the optimal parameters, and obtain the evaluation sub-model.

[0112] The present invention sorts out the quality evaluation indicators for the process pipeline installation, calculates its weight coefficients through a BP neural network, and finally obtains a scoring formula for the quality status of the process pipeline installation project to reflect the quality of the process pipeline installation project, which has a certain guiding role in the quality control and management of the process pipeline installation project. During the installation quality evaluation process, first use the evaluation sub-models of each first-level indicator constructed based on the BP neural network to calculate and evaluate the evaluation result data of the target second-level indicators corresponding to each first-level indicator to obtain the first-level evaluation results of the second-level indicators; then use the post-evaluation model to comprehensively evaluate the first-level evaluation results corresponding to each first-level indicator to evaluate the quality status of the process pipeline installation, which can comprehensively evaluate the quality of the process pipeline installation based on all quality evaluation indicators and has a high accuracy.

[0113] In an optional embodiment, for each first-level indicator in S204 above, based on the data set corresponding to the first-level indicator, using the grid search and cross-validation methods to train the BP neural network corresponding to the first-level indicator, determine the optimal parameters, and obtain the evaluation sub-model, includes:

[0114] S2041. For each first-level indicator, use the data set corresponding to the first-level indicator to train the corresponding BP neural network to determine the optimal number of nodes; wherein, the optimal number of nodes is the number of neurons in the hidden layer of the BP neural network.

[0115] In order to obtain the optimal number of nodes, it is necessary to experiment with different numbers of hidden layers and neurons, and determine the optimal number of nodes by comparing the mean square error values.

[0116] At the beginning, set the number of nodes to 2 and train the BP neural network. Then, save the value of the mean square error. Increase the number of nodes:

[0117] Increase the number of nodes to 2n, then retrain the neural network and record the mean square error.

[0118] Repeat increasing the number of nodes: Iterate the above process multiple times, each time increasing the number of nodes and retraining the neural network until the required maximum number of nodes is reached.

[0119] Compare the mean square errors: Calculate the mean square error corresponding to each number of nodes and record the results.

[0120] Select the optimal number of nodes: Based on the comparison results of the mean square errors, find the number of nodes with the minimum mean square error. This number of nodes m is the optimal number of nodes.

[0121] S2042. Update the BP neural network based on the optimal number of nodes to obtain an improved network model;

[0122] S2043. Iteratively train the improved network model based on the dataset corresponding to the first-level indicators, use grid search for parameter search, and use cross-validation to evaluate the performance of the model until the performance of the model meets the preset requirements to obtain the evaluation sub-model.

[0123] Values of the initial weight and the initial learning rate: Set the range of the initial learning rate as [0.1, 0.01, 0.001, 0.0001], use grid search and cross-validation to determine an optimal parameter value, and finally select the initial learning rate as 0.001; the initial weight is a random matrix, and the weight value is gradually trained and iterated according to the gradient descent method. The specific process is as follows:

[0124] (1) Use grid search and cross-validation to determine the optimal parameter:

[0125] (a) Initialize the range of grid search: Set the range of the initial learning rate as [0.1, 0.01, 0.001, 0.0001].

[0126] (b) Use grid search for parameter search: For each initial learning rate, use cross-validation to evaluate the performance of the model. Specifically, divide the dataset into multiple parts, use a part of the data for training, and then use another part of the data (i.e., the validation set) to evaluate the performance of the model. Repeat this process multiple times, each time using a different data splitting method to obtain a more accurate performance evaluation.

[0127] (c) Evaluate the performance of each parameter: For each initial learning rate, calculate some evaluation metrics such as mean squared error (MSE) or accuracy to understand how the model performs on the validation set.

[0128] (d) Select the optimal parameters: The present invention selects the initial learning rate and other parameters that perform best in cross-validation. This selection is based on evaluation metrics, such as selecting the learning rate with the smallest mean squared error.

[0129] (2) Use the gradient descent method to determine the weight values:

[0130] (a) Initialize the weight values: Generate a random matrix as the initial weight values. These initial weight values will be used for the training of the neural network.

[0131] (b) Define the loss function: Select the loss function (such as mean squared error) as the metric to measure the difference between the model prediction and the actual value.

[0132] (c) Calculate the gradient: Use the backpropagation algorithm to calculate the gradient of each weight with respect to the loss function. This will tell us how to adjust the weights to reduce the loss.

[0133] (d) Update the weights: According to the principle of the gradient descent method, update the weights by subtracting the value of the learning rate multiplied by the gradient from the current weights, which will gradually reduce the loss.

[0134] (e) Iterative training: Repeat steps (c) and (d) until the stopping condition is reached (such as reaching the maximum number of iterations or the loss function converges).

[0135] (f) Evaluate the model: After training is completed, use the validation set or test set to evaluate the performance of the model. Various metrics can be used to measure the accuracy and generalization ability of the model.

[0136] In an optional embodiment, the post-evaluation model is trained as follows:

[0137] After training the evaluation sub-models corresponding to each of the first-level indicators, divide the process pipeline installation project quality evaluation index database into a training set and a test set; according to the first-level indicators, input the training set into the corresponding evaluation sub-models to obtain the corresponding first-level evaluation result set; use the first-level evaluation result set to train the initial post-evaluation model, and use the gradient descent method to adjust the weights of each neuron in the input layer of the initial post-evaluation model to obtain the trained model; use the test set to perform cross-validation on the trained model, and repeat the process of adjusting the weights until the preset condition is reached to obtain the post-evaluation model.

[0138] Use the gradient descent method to determine the optimal weights of each first-level indicator, specifically as follows:

[0139] (a) Initialize weight values: Generate a random matrix as the initial weight values. These initial weight values will be used for the training of the neural network.

[0140] (b) Define the loss function: Select a loss function (such as mean squared error) as the metric to measure the difference between the model prediction and the actual value.

[0141] (c) Calculate the gradients: Use the backpropagation algorithm to calculate the gradients of each weight with respect to the loss function. This will tell us how to adjust the weights to reduce the loss.

[0142] (d) Update the weights: According to the principle of gradient descent, update the weights by subtracting the value of the learning rate multiplied by the gradient from the current weights. This will gradually reduce the loss.

[0143] (e) Iterative training: Repeat steps (c) and (d) until the stopping condition is reached (such as reaching the maximum number of iterations or the loss function converges).

[0144] (f) Evaluate the model: After the training is completed, use the validation set or the test set to evaluate the performance of the model. Various metrics can be used to measure the accuracy and generalization ability of the model. If the performance does not meet the preset requirements (such as being less than the preset error threshold), then repeat steps (c) - (d).

[0145] The above initial post - evaluation model includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is the number of the above - mentioned evaluation sub - models. The hidden layer performs a weighted sum of the first - level evaluation results of each evaluation sub - model input by the input layer to obtain the installation quality evaluation result, which is output through the output layer.

[0146] In an optional embodiment, collecting the first - hand data of the construction site, and screening out the evaluation index data related to the quality of the process pipeline installation project from the first - hand data, including:

[0147] Collect the first - hand data of the construction site, and screen out all the secondary index data related to the installation data quality control index, installation material quality control index, installation prefabrication quality control index, installation construction quality control index, and installation inspection quality control index from the first - hand data to obtain the evaluation index data.

[0148] In an optional embodiment, based on the pre - established process pipeline installation project quality evaluation index system table, screening out multiple target secondary indexes under each first - level index from the evaluation index data, including:

[0149] For the installation data quality control index, collect the data of design disclosure, drawing review, equipment unpacking inspection record, handover data, final acceptance, and construction plan submission for review from the evaluation index data;

[0150] For the quality control indicators of installation materials, collect the data of on-site assessment of welders, registration of qualified welders, submission for review of qualified welders and welding procedure qualification, compliance inspection of welding procedure qualification, and submission for review of engineering materials / equipment / fittings from the said evaluation index data;

[0151] For the quality control indicators of installation prefabrication, collect the data of process handover, heat treatment of prefabricated welds, hardness test of prefabricated pipelines, installation welding of prefabricated pipelines, and appearance inspection of welds from the said evaluation index data;

[0152] For the quality control indicators of installation construction, collect the data of heat treatment of pipeline installation welds, hardness of pipeline installation, non-destructive testing of pipeline installation, repair and extended exploration of pipeline installation welds, inspection of installation of pipeline compensation devices, installation of fixed / sliding pipe supports, adjustment of spring supports and hangers, installation of safety accessories, commissioning records of safety valves, welding records of pressure pipelines, static grounding test of pipelines, and installation data of pipelines connecting to machines from the said evaluation index data;

[0153] For the quality control indicators of installation inspection, collect the data of inspection of contact lines of metal ring gaskets / lens gaskets, closed water test of drainage pipelines, notification and supervision and inspection of pressure pipeline installation, technical disclosure, inspection records of welding consumables warehousing, temperature / humidity records of welding consumables warehouse, electrode drying records, electrode distribution and recovery records, inspection of materials entering the site, and on-site meteorological records from the said evaluation index data.

[0154] Sort out the quality evaluation indicators, calculate their weight coefficients through the BP neural network model, and finally obtain the scoring formula for the quality status of the process pipeline installation project to reflect the quality of the process pipeline installation project, which has a certain guiding role in the quality control and management of the process pipeline installation project.

[0155] In the embodiment of the present invention, the post-evaluation model is:

[0156] I = λ 1 A + λ 2 B + λ 3 C + λ 4 D + λ 5 E

[0157] In the formula, I is the scoring result of the installation project quality, λ 1 , λ 2 , λ 3 , λ 4 , λ 5Corresponding to the weight coefficients of each first-level indicator respectively, A is the first-level evaluation result of the installation data quality control indicator, B is the first-level evaluation result of the installation material quality control indicator, C is the first-level evaluation result of the installation prefabrication quality control indicator, D is the first-level evaluation result of the installation construction quality control indicator, and E is the first-level evaluation result of the installation inspection quality control indicator.

[0158] This application proposes a practical quantitative scoring method for the quality status evaluation of process pipeline installation in engineering construction. The evaluation process uses BP neural network to calculate the process pipeline installation quality evaluation index coefficient, and performs error analysis on the weight calculation value and theoretical value, and obtains good results, thereby proposing a process pipeline installation engineering quality scoring formula. This method not only takes into account the weight influencing factors of quality evaluation, but also saves all the information of evaluation at all levels, providing a certain reference for the evaluation of engineering construction quality.

[0159] Embodiment 2:

[0160] In order to more clearly explain the process pipeline installation quality evaluation method provided by the embodiment of the present invention and verify the accuracy of the method, the process pipeline installation quality evaluation method is applied to a certain petroleum refinery to obtain the installation quality evaluation result. The present application is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0161] Step 1: Conduct a survey on the construction unit in a certain oil refinery construction project, obtain first-hand data collected at the construction site through digital handover, clean the data to retain relevant data on the quality of the process pipeline installation project, and preliminarily screen the main influencing factors of the quality problems of the process pipeline installation: process pipeline installation data quality control, pipeline material quality control, prefabricated pipeline quality control, on-site installation quality control, and pipeline pressure test quality control. Among them, the process pipeline installation data quality control can be divided into drawing review and design disclosure, construction technology plan review, technical plan disclosure, welding process assessment and welding operation instruction inspection, welder qualification review, on-site machinery inspection of pipes, pipe fittings / flanges / valves quality certification documents and material accessories re-inspection, and welding material inspection according to unit projects;

[0162] Step 2: Since the data indicators for review and observation are somewhat subjective, to ensure the accuracy and effectiveness of the data, experts for this project are required to further confirm the main influencing factors of the quality problems in the installation of process pipelines and directly participate in the compilation of inspection and testing content and the classification of control point levels for different-level indicators. For example, for the secondary indicator of drawing review and design disclosure under the primary indicator of quality control of process pipeline installation data, the inspection and testing content is the drawing and design disclosure meeting, the control point level is A, and the inspection method is review, etc. Among them, each sub-project of the process pipeline installation project is used as the primary indicator, and the sub-indicator items of each sub-project are represented by A / B / C / D / E (i.e., the primary indicator), and each small item such as drawing review and design disclosure is a secondary indicator item, represented by A1 / A2 / A3, etc., thus generating a quality evaluation index system table for the process pipeline installation project;

[0163] Step 3: Based on the quality evaluation index system table for the process pipeline installation project, it serves as the quality evaluation index database for the process pipeline installation project.

[0164] Randomly select 50 groups of secondary indicator-related data from the data collected according to the quality evaluation index system table for the process pipeline installation project. Experts evaluate the indicator-related data to obtain the evaluation result data, and establish the corresponding quality evaluation index database for the process pipeline installation project. Among them, the control point level is used to represent the importance of the indicator. According to the four levels of control point levels A / B / C / D, they correspond to different scores of 4 / 3 / 2 / 1 respectively. Combining the number of levels of each secondary indicator item in the primary indicator item multiplied by the corresponding score can obtain the total score, and the weight value of each secondary indicator item is calculated and converted with 1 as the unit and input the data; taking the primary indicator containing 6 secondary indicator items as an example, including 3 A-level indicators, 2 B-level indicators, and 1 C-level indicator, then the weight of the A-level indicator is 100 / (4*3 + 3*2 + 2)*4% = 20%.

[0165] Step 4: Based on the BP neural network using the LM algorithm and combined with the quality evaluation index database for the process pipeline installation project, construct and train the improved network structure to obtain the process pipeline installation quality evaluation model, which can conveniently realize the quality evaluation of the process pipeline installation project. Specifically, it includes the following:

[0166] 1) Divide the quality evaluation index database for the process pipeline installation project into five data sets according to the first-level indicators. Each data set is divided into two parts: training samples and test samples. Analyzing the data shows that it is necessary to train the weights of five BP neural networks. The number of neurons in the input layer of each BP neural network is the number of secondary indicators within the corresponding primary indicator. The result a of the jth neuron in the lth layer [l] is:

[0167]

[0168] Among them, the superscript l represents the l-th layer of the neural network, a represents the neuron within the layer, σ is the activation function, is the weight coefficient of the k-th neuron on the l-th layer, b [l] is the bias coefficient of the l-th layer.

[0169] 2) Set network parameters. For the quality evaluation error analysis of process pipeline installation, obtain that the number of hidden layers is 1, and the number of neurons in the hidden layer is m, that is, m = 2 n , where n is any positive integer, and use the mean square error value for comparison.

[0170] To obtain the optimal number of nodes, we need to experiment with different numbers of hidden layers and neurons, and determine the optimal number of nodes by comparing the mean square error values. At the beginning, set the number of nodes to 2 and train the neural network. Then, save the value of the mean square error.

[0171] Increase the number of nodes: Increase the number of nodes, set it to 2n, and then retrain the neural network and record the mean square error.

[0172] Repeat increasing the number of nodes: Iterate the above process multiple times, each time increasing the number of nodes and retraining the neural network until the required maximum number of nodes is reached.

[0173] Compare the mean square errors: Calculate the mean square error corresponding to each number of nodes and record the results.

[0174] Select the optimal number of nodes: Based on the comparison results of the mean square errors, find the number of nodes with the smallest mean square error. This number of nodes m is the optimal number of nodes. In the embodiment of the present invention, the optimal number of nodes finally obtained is 8.

[0175] 3) Values of the initial weight and the initial learning rate: Set the range of the initial learning rate as [0.1, 0.01, 0.001, 0.0001], use grid search and cross-validation to determine an optimal parameter value, and finally select the initial learning rate as 0.001; the initial weight value is a random matrix, and the weight value is gradually trained and iterated according to the gradient descent method. The specific process is as follows:

[0176] (1) Use grid search and cross-validation to determine the optimal parameter

[0177] (a) Initialize the range of grid search: Set the range of the initial learning rate as [0.1, 0.01, 0.001, 0.0001].

[0178] (b) Parameter search using grid search: For each initial learning rate, cross-validation is used to evaluate the performance of the model. Specifically, we divide the dataset into multiple parts, use one part of the data for training, and then use another part of the data (i.e., the validation set) to evaluate the performance of the model. We repeat this process multiple times, each time using a different data splitting method to obtain a more accurate performance evaluation.

[0179] (c) Evaluate the performance of each parameter: For each initial learning rate, we will calculate some evaluation metrics, such as mean squared error (MSE) or accuracy, to understand the performance of the model on the validation set.

[0180] (d) Select the optimal parameters: We select the initial learning rate and other parameters that perform best in cross-validation. This selection is based on the evaluation metrics. For example, we may choose the learning rate with the smallest mean squared error.

[0181] (2) Determine the weight values using gradient descent:

[0182] (a) Initialize the weight values: Generate a random matrix as the initial weight values. These initial weight values will be used for the training of the neural network.

[0183] (b) Define the loss function: Select an appropriate loss function (such as mean squared error) as the metric to measure the difference between the model prediction and the actual value.

[0184] (c) Calculate the gradient: Use the backpropagation algorithm to calculate the gradient of each weight with respect to the loss function. This will tell us how to adjust the weights to reduce the loss.

[0185] (d) Update the weights: According to the principle of gradient descent, update the weights by subtracting the value of the learning rate multiplied by the gradient from the current weights. This will gradually reduce the loss.

[0186] (e) Iterative training: Repeat steps (c) and (d) until the stopping condition is reached (such as reaching the maximum number of iterations or the loss function converges).

[0187] (f) Evaluate the model: After training is completed, use the validation set or test set to evaluate the performance of the model. Various metrics can be used to measure the accuracy and generalization ability of the model.

[0188] 4) Precision optimization and computational speed optimization: Use mean squared error to judge. Compare the output result with the expected value. The smaller the error, the better the precision. For computational speed, use mini-batch gradient descent, which is faster than batch gradient descent and more accurate than stochastic gradient descent. Utilize the process pipeline installation quality evaluation database to divide the samples into A 1 -A 8 、B 1 -B6 , C 1 -C 6 , D 1 -D 6 , E 1 -E 6 Five groups of data are respectively input into the BP neural network to obtain five groups of indicators.

[0189] Among them, the weight table of the quality control indicators of the process pipeline installation data is shown in Table 2, and the error loss iteration analysis curve is as Figure 4 shown.

[0190]

[0191] Table 2

[0192] Using the quality evaluation database of the process pipeline installation, the true weight values of the first-level indicators of the process installation pipeline are shown in Table 3.

[0193] Table 3 Among them, λ 1 , λ 2 , λ 3 , λ 4 , λ 5 correspond to the weight coefficients for the control of the first-level indicators respectively. A is the first-level evaluation result of the quality control indicator of the installation data, B is the first-level evaluation result of the quality control indicator of the installation materials, C is the first-level evaluation result of the quality control indicator of the installation prefabrication, D is the first-level evaluation result of the quality control indicator of the installation construction, and E is the first-level evaluation result of the quality control indicator of the installation inspection. The weighted sum of the corresponding two is calculated using the following formula:

[0194] I = λ 1 A + λ 2 B + λ 3 C + λ 4 D + λ 5According to the provisions of the standard specifications and the actual requirements of the project, it can be known that when 85 < I ≤ 100, the quality evaluation is excellent; when 70 < I ≤ 85, the quality evaluation is good; when I = 70, the quality evaluation is qualified; when I < 70, the quality evaluation is unqualified. The above description shows that the BP neural network process pipeline installation quality evaluation method applied this time not only determines the main influencing factors of the process pipeline installation quality evaluation, but also provides a reference for the engineering quality evaluation method. This application proposes a practical quantitative scoring method for the evaluation of the quality status of process pipeline installation in engineering construction. In the evaluation process, the BP neural network is used to calculate the evaluation index coefficients of the process pipeline installation quality, and the error analysis is carried out on the calculated weight value and the theoretical value, and good results are obtained, thus proposing the quality scoring formula for the process pipeline installation project. This method not only considers the weight influencing factors in the process pipeline installation quality evaluation, but also preserves all the information of each level of evaluation, providing a certain reference for the evaluation related to the engineering construction quality. Example 3 The embodiment of the present invention provides a process pipeline installation quality evaluation device, as shown in Figure 5 shown, including: a collection module 301, configured to collect first-hand data at the construction site, and screen out evaluation index data related to the quality of the process pipeline installation project from the first-hand data; a screening module 302, configured to screen out multiple target secondary indicators under each first-level indicator from the evaluation index data based on a pre-established process pipeline installation project quality evaluation index system table; an evaluation module 303, configured to obtain the evaluation result data of each target secondary indicator under each first-level indicator, and input the evaluation result data under each first-level indicator into the process pipeline installation quality evaluation model to obtain the installation quality scoring result.

[0195] The process pipeline installation quality evaluation device provided by the embodiment of the present invention has the same implementation principle and technical effect as any one of the foregoing method embodiments, and will not be described in detail here.

[0196] Example 4

[0197] The embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the process pipeline installation quality evaluation method as described in any one of the foregoing method embodiments.

[0198] This computer-readable storage medium may be included in the device / apparatus described in the above embodiments; or it may exist alone without being assembled into the device / apparatus. The above computer-readable storage medium carries one or more programs, and when the one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0199] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present invention, the computer-readable storage medium may be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0200] Embodiment Five

[0201] An embodiment of the present invention provides an electronic device. Referring to Figure 6 as shown, it includes a processor 111, a communication interface 112, a memory 113, and a communication bus 114. Among them, the processor 111, the communication interface 112, and the memory 113 complete mutual communication through the communication bus 114.

[0202] The memory 113 is used to store a computer program.

[0203] When the processor 111 is used to execute the program stored on the memory 113, it implements the process pipeline installation quality evaluation method described in any of the foregoing method embodiments.

[0204] For the electronic device provided by the embodiment of the present invention, its implementation principle and technical effects are similar to those of any of the foregoing method embodiments, and will not be elaborated here.

[0205] The above-mentioned memory 113 may be an electronic memory such as flash memory, EEPROM (electrically erasable programmable read-only memory), EPROM, hard disk, or ROM. The memory 113 has a storage space for program codes for executing any method steps in the above methods. For example, the storage space for program codes may include respective program codes for implementing each step in the above methods. These program codes can be read from or written into one or more computer program products. These computer program products include program code carriers such as hard disks, optical discs (CDs), memory cards, or floppy disks. Such computer program products are usually portable or fixed storage units. The storage unit may have a storage segment or storage space arranged similarly to the memory 113 in the above-mentioned electronic device. The program codes may be compressed in an appropriate form, for example. Generally, the storage unit includes a program for executing the method steps according to the embodiments of the present invention, that is, codes that can be read by, for example, the processor 111. When these codes are run by the electronic device, they cause the electronic device to execute each step in the method described above.

[0206] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other, and the key point of each embodiment is to illustrate the differences from other embodiments. It should be noted that, without conflict, the embodiments and the features in the embodiments of the present invention can be combined with each other. The present invention is not limited to any single aspect, nor to any single embodiment, nor to any arbitrary combination and / or permutation of these aspects and / or embodiments. Each aspect and / or embodiment of the present invention can be used alone, or in combination with one or more other aspects and / or other embodiments.

[0207] Finally, it should be noted that the above-described embodiments are only specific embodiments of the present invention, used to illustrate the technical solutions of the present invention, rather than to limit it. The protection scope of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that any person skilled in the art within the technical scope disclosed by the present invention can still modify the technical solutions recorded in the foregoing embodiments, or can easily think of changes, or perform equivalent replacements on some of the technical features; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for evaluating the installation quality of process pipelines, characterized in that, it includes: Collect first-hand data at the construction site, and screen out evaluation index data related to the installation quality of process pipelines from the first-hand data; Based on the pre-established quality evaluation index system table for process pipeline installation projects, screen out multiple target secondary indicators under each first-level indicator from the evaluation index data; Obtain the evaluation result data of each of the target secondary indicators under each first-level indicator, and input the evaluation result data under each first-level indicator into the process pipeline installation quality evaluation model to obtain the installation quality score result.

2. The method for evaluating the installation quality of process pipelines according to claim 1, characterized in that, The process pipeline installation quality evaluation model includes a post-evaluation model and evaluation sub-models corresponding to each first-level indicator; The step of inputting the evaluation result data under each first-level indicator into the process pipeline installation quality evaluation model to obtain the installation quality score result includes: For each first-level indicator, input the evaluation result data under the first-level indicator into the evaluation sub-model corresponding to the first-level indicator to obtain the corresponding first-level evaluation result; Input the first-level evaluation results corresponding to each first-level indicator into the post-evaluation model to obtain the installation quality score result.

3. The method for evaluating the installation quality of process pipelines according to claim 2, characterized in that, The evaluation sub-model is trained in the following manner: Obtain a quality evaluation index database for process pipeline installation projects; For each first-level indicator in the quality evaluation index database for process pipeline installation projects, establish a corresponding BP neural network respectively; wherein, the number of neurons in the input layer of the BP neural network is the number of secondary indicators under the corresponding first-level indicator; Divide the quality evaluation index database for process pipeline installation projects according to the first-level indicators to obtain a data set corresponding to each first-level indicator; For each first-level indicator, based on the data set corresponding to the first-level indicator, use the grid search and cross-validation methods to train the BP neural network corresponding to the first-level indicator, determine the optimal parameters, and obtain the evaluation sub-model.

4. The method for evaluating the installation quality of process pipelines according to claim 3, characterized in that, The step of, for each first-level indicator, based on the data set corresponding to the first-level indicator, using the grid search and cross-validation methods to train the BP neural network corresponding to the first-level indicator, determine the optimal parameters, and obtain the evaluation sub-model includes: For each first-level indicator, use the data set corresponding to the first-level indicator to train the corresponding BP neural network to determine the optimal number of nodes; wherein, the optimal number of nodes is the number of neurons in the hidden layer of the BP neural network; Update the BP neural network based on the optimal number of nodes to obtain an improved network model; Iteratively train the improved network model based on the data set corresponding to the first-level indicator, use grid search for parameter search, and use cross-validation to evaluate the performance of the model until the performance of the model meets the preset requirements to obtain the evaluation sub-model.

5. The process pipeline installation quality evaluation method according to claim 2, characterized in that, the post-evaluation model is trained through the following method: After training the evaluation sub-models corresponding to each of the first-level indicators, divide the process pipeline installation project quality evaluation index database into a training set and a test set; According to the first-level indicators, input the training set into the corresponding evaluation sub-models respectively to obtain the corresponding first-level evaluation result sets; Use the first-level evaluation result sets to train the initial post-evaluation model, and use the gradient descent method to adjust the weights of each neuron in the input layer of the initial post-evaluation model to obtain the trained model; Use the test set to perform cross-validation on the trained model, and repeat the process of adjusting the weights until the preset conditions are met to obtain the post-evaluation model.

6. The process pipeline installation quality evaluation method according to claim 2, characterized in that, inputting the first-level evaluation results corresponding to each first-level indicator into the post-evaluation model to obtain the installation quality score result, including: inputting the first-level evaluation results corresponding to each first-level indicator into the post-evaluation model as follows to obtain the installation quality score result: I = λ 1 X 1 +...+ λ i X i +...+ λ n X n where I is the installation quality scoring result, and λ 1 ,..., λ i ,..., λ n are the weight coefficients of each first-level indicator respectively, and X 1 ,..., X i ,..., X n are the first-level evaluation results of each first-level indicator respectively.

7. The process pipeline installation quality evaluation method according to claim 1, characterized in that, collecting the first-hand data at the construction site, and screening out the evaluation index data related to the process pipeline installation project quality from the first-hand data, including: collecting the first-hand data at the construction site, and screening out all the secondary index data related to the installation data quality control index, installation material quality control index, installation prefabrication quality control index, installation construction quality control index and installation inspection quality control index from the first-hand data to obtain the evaluation index data.

8. The process pipeline installation quality evaluation method according to claim 7, characterized in that, based on the pre-established process pipeline installation project quality evaluation index system table, screening out multiple target secondary indicators under each first-level indicator from the evaluation index data, including: For the installation data quality control index, collecting the design disclosure, drawing review, equipment unpacking inspection record, handing-over data, final acceptance, and construction plan approval data from the evaluation index data; For the installation material quality control index, collecting the on-site welder assessment, registration of qualified welders, submission of qualified welders and welding procedure qualification, compliance inspection of welding procedure qualification, and submission of engineering materials / equipment / components from the evaluation index data; For the installation prefabrication quality control index, collecting the process handover, heat treatment of prefabricated welds, hardness test of prefabricated pipelines, installation welding of prefabricated pipelines and weld appearance inspection data from the evaluation index data; For the installation construction quality control index, collecting the heat treatment of pipeline installation welds, pipeline installation hardness, non-destructive testing of pipeline installation, repair and re-inspection of pipeline installation welds, inspection of pipeline compensation device installation, installation of fixed / sliding pipe supports, adjustment of spring supports and hangers, installation of safety accessories, safety valve commissioning record, pressure pipeline welding record, pipeline static grounding test, and installation data of pipelines connecting to machines from the evaluation index data; For the quality control indicators of installation inspection, collect the inspection records of the contact lines of metal ring gaskets / lens gaskets, the water tightness test of drainage pipes, the notification and supervision inspection of the installation of pressure pipes, the technical disclosure, the inspection records of welding consumables entering the warehouse, the temperature / humidity records of the welding consumables warehouse, the drying records of welding electrodes, the issue and return records of welding electrodes, the inspection of materials entering the site, and the on-site meteorological record data from the evaluation index data.

9. According to the process pipeline installation quality evaluation method described in claim 1, characterized in that, the method further includes: According to the pre-established quality specifications, conduct a grade evaluation on the installation quality scoring result to obtain a quality rating result.

10. A process pipeline installation quality evaluation device, characterized in that, comprising: A collection module, configured to collect first-hand data at the construction site, and screen out evaluation index data related to the quality of the process pipeline installation project from the first-hand data; A screening module, configured to screen out multiple target secondary indicators under each primary indicator from the evaluation index data based on the pre-established process pipeline installation project quality evaluation index system table; An evaluation module, configured to obtain the evaluation result data of each target secondary indicator under each primary indicator, and input the evaluation result data under each primary indicator into the process pipeline installation quality evaluation model to obtain an installation quality scoring result.

11. A computer-readable storage medium, on which a computer program is stored, characterized in that, When the program is executed by a processor, it implements the process pipeline installation quality evaluation method described in any one of claims 1-9.

12. An electronic device, characterized in that, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory complete communication with each other through the communication bus; The memory is used to store a computer program; The processor is configured to implement the process pipeline installation quality evaluation method described in any one of claims 1-9 when executing the program stored on the memory.

Images