Method and device for testing explosion-proof performance of explosion-proof diesel engine system based on intelligent sensing

Abstract

The invention discloses an explosion-proof diesel engine system explosion-proof performance test method and device based on intelligent sensing, and relates to the technical field of explosion-proof diesel engine tests.The method comprises the steps that multi-source sensing data of an explosion-proof diesel engine in the operation process are obtained, and the multi-source sensing data comprise the temperature, the pressure and the combustible gas concentration; based on the multi-source sensing data, constructing an explosion-proof performance evaluation model, and outputting an explosion-proof performance score value; if the anti-explosion performance score value is lower than a preset safety threshold value, an operation parameter adjusting strategy is generated, and operation parameters of the diesel engine are dynamically adjusted; and re-collecting the adjusted multi-source sensing data, verifying whether the anti-explosion performance score value reaches a preset safety threshold value or not, and if so, judging that the anti-explosion performance is qualified. The problems that an existing explosion-proof diesel engine system explosion-proof performance testing method only pays attention to single parameter threshold value judgment, and parameter adjustment depends on an empirical fixed strategy, so that the explosion-proof performance evaluation precision is low, the working condition adaptability is poor, and the testing efficiency and reliability are insufficient are solved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of explosion-proof diesel engine test, in particular to an explosion-proof performance test method and device for an explosion-proof diesel engine system based on intelligent sensing. BACKGROUND

[0002] With the rapid development of industry, explosion-proof diesel engines are increasingly widely used in dangerous environments such as coal mines and chemical plants, and accurate evaluation and effective protection of explosion-proof performance are increasingly important. However, the traditional explosion-proof performance test method for the explosion-proof diesel engine system has the problems of low evaluation accuracy of explosion-proof performance, poor working condition adaptability, insufficient test efficiency and reliability, and the like, due to the fact that it only focuses on single parameter threshold determination and parameter adjustment relies on empirical fixed strategies. In addition, the method lacks deep mining and comprehensive utilization of multi-source sensing data, and cannot provide effective decision basis for production and maintenance in time. SUMMARY

[0003] The embodiments of the application provide an explosion-proof performance test method and device for an explosion-proof diesel engine system based on intelligent sensing, and solve the technical problem that the existing explosion-proof performance test method for the explosion-proof diesel engine system only focuses on single parameter threshold determination, and parameter adjustment relies on empirical fixed strategies, resulting in low evaluation accuracy of explosion-proof performance, poor working condition adaptability, and insufficient test efficiency and reliability.

[0004] The technical solution of the application to solve the above technical problem is as follows: In a first aspect, the application provides an explosion-proof performance test method for an explosion-proof diesel engine system based on intelligent sensing, which comprises: acquiring multi-source sensing data of the explosion-proof diesel engine during operation, wherein the multi-source sensing data at least includes temperature, pressure and combustible gas concentration; constructing an explosion-proof performance evaluation model based on the multi-source sensing data, and outputting an explosion-proof performance score value; if the explosion-proof performance score value is lower than a preset safety threshold, generating an operation parameter adjustment strategy to dynamically adjust the operation parameters of the diesel engine; reacquiring the adjusted multi-source sensing data, verifying whether the explosion-proof performance score value reaches the preset safety threshold, and if so, determining that the explosion-proof performance is qualified.

[0005] In a second aspect, the application provides an explosion-proof performance test device for an explosion-proof diesel engine system based on intelligent sensing, which comprises: a data acquisition module for acquiring multi-source sensing data of the explosion-proof diesel engine during operation, wherein the multi-source sensing data at least includes temperature, pressure and combustible gas concentration; a model construction module for constructing an explosion-proof performance evaluation model based on the multi-source sensing data, and outputting an explosion-proof performance score value; The parameter adjustment module is configured to generate a running parameter adjustment strategy to dynamically adjust the running parameters of the diesel engine if the explosion-proof performance score value is lower than the preset safety threshold. The performance determination module is configured to re-collect the adjusted multi-source sensing data, verify whether the explosion-proof performance score value reaches the preset safety threshold, and determine that the explosion-proof performance is qualified if the explosion-proof performance score value reaches the preset safety threshold.

[0006] The present application provides one or more technical solutions, at least having the following technical effects or advantages: The explosion-proof diesel engine system explosion-proof performance test method and device based on intelligent sensing provided by the present application first use the multi-source sensing data of the explosion-proof diesel engine in the running process to evaluate the explosion-proof performance of the explosion-proof diesel engine system. Secondly, a scientific explosion-proof performance evaluation model is constructed based on the multi-source sensing data, avoiding the limitation of only focusing on single parameter threshold determination, and improving the accuracy of explosion-proof performance evaluation. At the same time, a dynamic running parameter adjustment strategy is generated according to real-time multi-source sensing data, overcoming the problem of relying on experiential fixed strategy for parameter adjustment, and enhancing the adaptability of the system to different working conditions.

[0007] Through the above technical solutions, accurate evaluation and effective adjustment of the explosion-proof performance of the explosion-proof diesel engine system can be realized, and the test efficiency and reliability are improved. Deep mining and comprehensive utilization of multi-source sensing data can timely discover potential safety hazards, provide effective decision basis for production and maintenance, and reduce the risk of safety accidents. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0009] Figure 1 is a flow diagram of an explosion-proof performance test method for an explosion-proof diesel engine system based on intelligent sensing provided by the present application; Figure 2 is a structural schematic diagram of an explosion-proof performance test device for an explosion-proof diesel engine system based on intelligent sensing provided by the present application; Figure 3 is a structural schematic diagram of an intelligent explosion-proof diesel engine system engine provided by the present application; Figure 4 is a test data table of an intelligent explosion-proof diesel engine system provided by the present application.

[0010] In the drawings, the components represented by the numbers are described as follows: Data acquisition module 11, model construction module 12, parameter adjustment module 13, performance determination module 14, radiator assembly 111, fan 112, air intake flame arrestor assembly 113, water-cooled supercharger assembly 114, water washing tank assembly 115, fuel tank assembly 116. DETAILED DESCRIPTION

[0011] The embodiment of the present application provides a kind of explosion-proof diesel engine system explosion-proof performance test method and device based on intelligent sensing, for solving the technical problem that existing explosion-proof diesel engine system explosion-proof performance test method only pays attention to single parameter threshold determination, and parameter adjustment relies on experience fixed strategy, leading to low precision of explosion-proof performance evaluation, poor working condition adaptability, insufficient test efficiency and reliability.

[0012] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0013] In the description of the present application, the terms "first", "second" are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined with "first", "second" can be explicitly or implicitly included one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0014] In the description of the present application, the term "for example" is used to indicate "as an example, illustration or explanation". Any embodiment described as "for example" in the present application is not necessarily interpreted as more preferred or more advantageous than other embodiments. In order to enable any person skilled in the art to implement and use the present application, the following description is given. In the following description, details are listed for the purpose of explanation. It should be understood that those skilled in the art can realize the present application without using these specific details. In other examples, well-known structures and processes will not be described in detail to avoid unnecessary details making the description of the present application obscure. Therefore, the present application is not intended to be limited to the shown embodiments, but is consistent with the broadest scope that meets the principles and characteristics disclosed in the present application.

[0015] Embodiment one, as Figure 1 shown, the embodiment of the present application provides an explosion-proof diesel engine system explosion-proof performance test method based on intelligent sensing, comprising: S10: Obtain multi-source sensing data of the explosion-proof diesel engine during operation, wherein the multi-source sensing data at least includes temperature, pressure and combustible gas concentration; In the embodiments of the present application, first, the data of the explosion-proof diesel engine during operation is collected in real time by multi-source sensors, and the multi-source sensing data includes temperature, pressure and combustible gas concentration. The temperature sensor measures the temperature change of the key parts of the diesel engine, the pressure sensor measures the pressure fluctuation in the system, and the combustible gas concentration sensor measures the content of combustible gas in the environment.

[0016] Among them, temperature, pressure and combustible gas concentration are factors affecting the explosion-proof performance of the explosion-proof diesel engine system, which jointly determine the safety and stability of the explosion-proof diesel engine during operation. Abnormal rise of temperature may cause combustible gas to reach ignition point, increasing the risk of explosion; sudden change of pressure may cause system failure, affecting the normal operation of the system; too high concentration of combustible gas directly constitutes a dangerous condition for explosion.

[0017] The data collection frequency of the multi-source sensor is set according to the actual demand and system characteristics, to ensure that the obtained data is timely and accurate. At the same time, in order to ensure the data quality, the collected raw data is preliminarily preprocessed, such as removing noise, etc., to provide data basis for subsequent data analysis and model construction.

[0018] Specifically, step S10 in the method includes: By arranging intelligent sensor nodes at the exhaust port, air inlet and engine compartment of the explosion-proof diesel engine, temperature, pressure and combustible gas concentration data are collected in real time; The collected data is time-aligned and noise-filtered to form a standardized sensing data sequence.

[0019] In the embodiments of the present application, first, intelligent sensor nodes are arranged at the exhaust port, air inlet and engine compartment of the explosion-proof diesel engine to collect data in real time. The collected data includes key data such as collected temperature, pressure and combustible gas concentration, etc. Figure 4 Parameter categories.

[0020] Among them, the sensor arranged at the exhaust port can accurately monitor the temperature, pressure and combustible gas concentration of the exhaust emission, reflecting the degree of combustion and exhaust emission; the sensor at the air inlet can obtain the air parameters entering the diesel engine in real time, providing basis for adjusting the air intake and fuel injection amount; the sensor in the engine compartment can monitor the internal operating environment of the engine to ensure that each component works under suitable temperature and pressure conditions.

[0021] For example, Figure 3As shown, the structure diagram of the explosion-proof diesel engine in the embodiment of the present application includes a radiator assembly 111, a fan 112, an air intake flame arrest assembly 113, a water-cooled supercharger assembly 114, a water washing box assembly 115, and a fuel tank assembly 116, which are respectively used for heat dissipation, flame arrest, and exhaust temperature reduction to achieve explosion-proof.

[0022] Secondly, after collecting the data, time alignment and noise filtering processing are performed. Due to the possible slight difference in sampling time of different sensors, time alignment can make the data of each sensor consistent in the time dimension, facilitating subsequent analysis and processing. Noise filtering processing is to remove the noise in the data caused by external interference or sensor errors, improving the quality and reliability of the data.

[0023] For example, common noise filtering methods include mean filtering, median filtering, Kalman filtering, etc. The embodiment adopts the mean filtering method, which smoothes the data by calculating the average value of the data within a certain window to reduce the influence of noise. In specific operation, a certain length window is selected, for example, a window of 5 data points, and the average value of the data in the window is calculated as the filtered value of the point. After mean filtering processing, the fluctuation of the data is reduced, and the real physical quantity change can be better reflected.

[0024] The standardized sensor data sequence formed after time alignment and noise filtering processing has a unified format and scale, which is more conducive to subsequent feature extraction and model construction.

[0025] S20: Based on the multi-source sensor data, an explosion-proof performance evaluation model is constructed, and an explosion-proof performance score value is output. In the embodiment of the present application, based on the collected multi-source sensor data, a machine learning algorithm is used to construct an explosion-proof performance evaluation model. Specifically, the standardized sensor data sequence is taken as input, and a neural network is used to construct and train the explosion-proof performance evaluation model.

[0026] Further, taking a neural network as an example, a multi-layer perceptron including an input layer, a hidden layer, and an output layer is constructed. The input layer receives temperature, pressure, and combustible gas concentration multi-source sensor data, the hidden layer performs nonlinear transformation and feature extraction on the input data through multiple neurons, and the output layer outputs the explosion-proof performance score value.

[0027] During the training process, a large amount of historical data is used for supervised learning. The historical data contains multi-source sensor data under different working conditions and corresponding real explosion-proof performance evaluation results. By continuously adjusting the weights and biases of the neural network, the output of the model is as close as possible to the real evaluation result. After training and optimization, the explosion-proof performance evaluation model can accurately output the explosion-proof performance score value according to real-time multi-source sensor data.

[0028] Specifically, step S20 in the method comprises: extracting a feature vector of the multi-source sensing data; inputting the feature vector into a pre-trained explosion-proof performance evaluation model to output an explosion-proof performance score value, wherein the higher the explosion-proof performance score value is, the better the explosion-proof performance is.

[0029] In the embodiments of the present application, first, since the original multi-source sensing data often has high dimensions and contains a large amount of redundant information, directly using the original data for model training will increase the computational complexity and training time, and may even cause model overfitting, so a feature vector is extracted from the multi-source sensing data.

[0030] Secondly, after the feature vector is extracted, it is input into a pre-trained explosion-proof performance evaluation model. The explosion-proof performance evaluation model learns the mapping relationship between the feature vector and the true explosion-proof performance evaluation result in the historical data, evaluates the new feature vector, and outputs an explosion-proof performance score value. The higher the explosion-proof performance score value is, the better the explosion-proof performance of the explosion-proof diesel engine system is, and the higher the safety and stability of the system in the running process are.

[0031] Among them, extracting the feature vector of the multi-source sensing data comprises: segmenting the standardized sensing data sequence according to a preset sliding time window; for the temperature, pressure and combustible gas concentration data in each time window, the corresponding time domain statistical features including the maximum value, the minimum value, the mean value and the standard deviation are calculated respectively; performing frequency domain transformation on the data in each time window, and extracting the amplitude of the dominant frequency component as the frequency domain feature; combining the time domain statistical features and the frequency domain features, and adaptively weighting and fusing based on the historical variation degree of each sensing data to form the final feature vector.

[0032] In the embodiments of the present application, first, the standardized sensing data sequence is segmented according to a preset sliding time window. That is, the temperature, pressure and combustible gas concentration data collected by the sensor and after being arranged are divided according to the fixed time interval set in advance, such as 1 minute, to divide the continuous temperature, pressure and combustible gas concentration data.

[0033] Secondly, for the temperature, pressure and combustible gas concentration data, the mean value, variance, maximum value, minimum value and change rate can be extracted. The mean value reflects the average level of the temperature, and the variance reflects the fluctuation of the temperature. The maximum value and the minimum value can help to judge whether the temperature is out of the safe range, and the change rate can capture the dynamic change trend of the temperature.

[0034] Further, the maximum value, the minimum value, the average value and the fluctuation amplitude standard deviation of the temperature are calculated respectively, and the pressure and the combustible gas concentration are also calculated in the same way as time domain statistical features.

[0035] Again, each piece of data is converted into frequency fluctuation information, and the most critical fluctuation frequency and corresponding intensity are found as frequency domain features reflecting the dynamic situation of the data.

[0036] Finally, the time domain statistical features and the frequency domain features of the same time period are integrated, and according to the fluctuation degree of the temperature, pressure and combustible gas concentration data in the past normal operation, different features are assigned different importance weights, and the weighted comprehensive feature data are obtained for evaluation.

[0037] Exemplarily, assuming that the preset sliding time window is 1 minute, the standardized data of 10 minutes is divided into 10 segments, and one piece of data is taken: the temperature data is 35℃, 38℃, 36℃, 40℃, 37℃, 39℃, and the time domain statistical features are calculated as the maximum value 40℃, the minimum value 35℃, the average value 37.5℃, and the standard deviation 1.8℃; The pressure data is 0.3MPa, 0.5MPa, 0.4MPa, 0.6MPa, 0.4MPa, 0.5MPa, and the time domain statistical features are the maximum value 0.6MPa, the minimum value 0.3MPa, the average value 0.45MPa, and the standard deviation 0.11MPa; The combustible gas concentration data is 0.04%, 0.06%, 0.05%, 0.07%, 0.05%, 0.06%, and the time domain statistical features are the maximum value 0.07%, the minimum value 0.04%, the average value 0.055%, and the standard deviation 0.011%.

[0038] Further, the data in each time window is subjected to multi-frequency domain transformation to obtain the amplitude of the dominant frequency component.

[0039] Among them, the data in each time window is subjected to frequency domain transformation, and the amplitude of the dominant frequency component is extracted as the frequency domain feature, including: The fast Fourier transform is performed on the sensor data sequence in each time window, the sensor data sequence is converted from time domain to frequency domain, and the corresponding frequency domain amplitude sequence is obtained; The total energy of all frequency components in the frequency domain amplitude sequence is calculated; According to the amplitude size, the frequency components are sorted, and the energy is accumulated from the highest amplitude component until the accumulated energy reaches the preset percentage threshold of the total energy; The frequency component corresponding to the accumulated energy is determined as the dominant frequency component, and the corresponding amplitude is recorded as the frequency domain feature.

[0040] In the embodiments of the present application, firstly, the sensor data sequence in each time window is converted into data of temperature fluctuation intensity at different frequencies by fast Fourier transform, so as to obtain a fluctuation intensity sequence corresponding to each frequency.

[0041] The Fourier transform is a prior art means for converting a signal between time domain and frequency domain, and converts the complex sensor data changing with time into fluctuation intensity data at different frequencies, so as to analyze the frequency characteristics of the data.

[0042] Further, the principle of Fourier transform is based on that any periodic signal can be expressed as a superposition of a series of sine and cosine functions. In the present application, the sensor data sequence in each time window is subjected to fast Fourier transform, and the running state of the system is described in combination with the time domain statistical characteristics.

[0043] Secondly, the total fluctuation amount of the data, i.e. the total energy, is calculated by cumulatively adding the fluctuation intensities of all frequencies.

[0044] Then, the frequencies are sorted in descending order of fluctuation intensity, and the fluctuation intensities are cumulatively added from the strongest frequency, until the cumulative amount reaches a preset proportion of the total energy, such as 80% as set in advance.

[0045] Finally, the frequency component corresponding to the frequency whose cumulative energy reaches the preset proportion of the total energy is determined as the dominant frequency component, and the corresponding fluctuation intensity is recorded as the frequency domain feature.

[0046] For example, it is assumed that in a 1-minute time window, the sensor records the pressure data of the exhaust port of the explosion-proof diesel engine every 10 seconds, and obtains 6 data: 0.3 MPa, 0.5 MPa, 0.4 MPa, 0.6 MPa, 0.4 MPa, and 0.5 MPa. The data is subjected to fast Fourier transform to obtain a frequency domain amplitude sequence of 4 corresponding to 0.5 times per second, 6 corresponding to 1 time per second, and 2 corresponding to 1.5 times per second. The total energy is calculated to be 4+6+2=12, and the preset percentage threshold is 80%, i.e. 9.6. After sorting in descending order of fluctuation intensity, the fluctuation intensity 6 corresponding to 1 time per second is added first, and the cumulative energy 6 does not reach 9.6. Then, the fluctuation intensity 4 corresponding to 0.5 times per second is added, and the cumulative energy 10 reaches 9.6. Therefore, 1 time per second and 0.5 times per second are determined as the key frequencies, and the fluctuation intensities 6 and 4 are recorded as the frequency domain features.

[0047] Further, the time domain statistical characteristics and the frequency domain features are combined, and adaptively weighted and fused based on the historical variation degree of each sensor data to form a final feature vector, which includes: The time domain statistical features and the frequency domain features extracted from the temperature, pressure and combustible gas concentration data are spliced in order of sensor types to form an initial fusion feature vector; A coefficient of variation of each sensor data under a historical normal operation state is calculated, wherein the coefficient of variation is a ratio of a standard deviation to a mean value; The reciprocal of the coefficient of variation is taken as a basic value of each feature weight, and an adaptive weight coefficient corresponding to each feature is obtained through normalization processing; A diagonal weight matrix with the adaptive weight coefficients as diagonal elements is constructed, and the initial fusion feature vector is multiplied by the diagonal weight matrix to obtain a final feature vector after weighting.

[0048] In the embodiments of the present application, the time domain statistical features and the frequency domain features of the temperature, pressure and combustible gas concentration obtained above are integrated, and then different weights are assigned to the features according to the historical stability of each type of data according to the importance, to finally form comprehensive feature data that can accurately evaluate the explosion-proof performance.

[0049] Specifically, first, the time domain statistical features and the frequency domain features of each type of data are spliced in order of temperature features-pressure features-combustible gas concentration features to form an initial comprehensive feature, i.e., an initial fusion feature vector, which covers the key information of each sensor data in the time domain and the frequency domain, providing a basis for subsequent weighted fusion.

[0050] Secondly, the coefficient of variation of each type of sensor data under past normal operation is calculated. The coefficient of variation is an important indicator for measuring the degree of data dispersion. By calculating the coefficient of variation, the fluctuation of each sensor data under normal conditions is understood. The greater the coefficient, the more unstable the data, and the greater the impact on the explosion-proof performance.

[0051] Then, since the smaller the coefficient of variation, the more stable the data, the corresponding feature should have a higher weight when evaluating the explosion-proof performance. The reciprocal of the coefficient of variation is taken as the basic value of each feature weight, and normalization processing is used to standardize the data to obtain the adaptive weight of each feature. Normalization processing can ensure that the sum of the weights of all features is 1, avoiding affecting the accuracy of the evaluation result due to excessively large or small weights.

[0052] Finally, a diagonal matrix is constructed with the adaptive weight coefficients as diagonal elements, and the initial fusion feature vector is multiplied by the diagonal matrix to obtain a final comprehensive feature vector after weighting. The diagonal matrix has the characteristic that all elements except those on the diagonal line are 0. The final feature vector considers the time domain and frequency domain features of each sensor data and is weighted according to the historical stability of the data, which can more accurately reflect the running state and explosion-proof performance of the explosion-proof diesel engine system.

[0053] Exemplarily, assuming that the time domain feature of the temperature data is the maximum value 50℃ and the average value 42℃, and the frequency domain feature is the key frequency fluctuation intensity 6 and 4; the time domain feature of the pressure data is the maximum value 0.6MPa and the average value 0.4MPa, and the frequency domain feature is the key frequency fluctuation intensity 6 and 4; the time domain feature of the combustible gas concentration data is the maximum value 0.08% and the average value 0.05%, and the frequency domain feature is the key frequency fluctuation intensity 3 and 2, the initial fusion feature vector is spliced in order as [50, 42, 6, 4, 0.6, 0.4, 6, 4, 0.08, 0.05, 3, 2].

[0054] The coefficient of variation in the historical normal state is calculated, the temperature coefficient of variation is 0.15, which is relatively unstable; the pressure coefficient of variation is 0.1, which is relatively stable; the coefficient of variation of the combustible gas concentration is 0.2, which is the least stable. The reciprocals of the coefficients of variation are 6.67, 10 and 5 respectively, and the normalized temperature feature weight is 0.32, the pressure feature weight is 0.48, and the combustible gas concentration feature weight is 0.2.

[0055] Further, the diagonal weight matrix is constructed and multiplied by the initial vector to obtain the final feature vector.

[0056] Exemplarily, the diagonal weight matrix is constructed, and the initial fusion feature vector is [50, 42, 6, 4, 0.6, 0.4, 6, 4, 0.08, 0.05, 3, 2]; after calculation, the adaptive weight coefficients corresponding to each feature are [0.15, 0.15, 0.1, 0.1, 0.2, 0.2, 0.05, 0.05, 0.03, 0.03, 0.02, 0.02] in turn.

[0057] The diagonal weight matrix is constructed, which is a 12x12 square matrix, only the elements on the diagonal are the above-mentioned adaptive weight coefficients, and the elements in the remaining positions are all 0, that is, the first to twelfth elements on the diagonal are 0.15, 0.15, 0.1, 0.1, 0.2, 0.2, 0.05, 0.05, 0.03, 0.03, 0.02, 0.02 in turn.

[0058] Finally, the initial fusion feature vector, i.e. the 1x12 vector, is multiplied by the diagonal weight matrix, each element in the vector is multiplied by the diagonal weight coefficient in the corresponding position of the matrix during calculation, and the final feature vector after weighting is obtained: [50x0.15, 42x0.15, 6x0.1, 4x0.1, 0.6x0.2, 0.4x0.2, 6x0.05, 4x0.05, 0.08x0.03, 0.05x0.03, 3x0.02, 2x0.02], i.e. [7.5, 6.3, 0.6, 0.4, 0.12, 0.08, 0.3, 0.2, 0.0024, 0.0015, 0.06, 0.04], which is the final feature data of the explosion-proof performance evaluation model.

[0059] Further, the training step of the explosion-proof performance evaluation model comprises: extracting a sample feature vector set of the historical operation multi-source sensor data of the explosion-proof diesel engine, and collecting the true value of the explosion-proof performance score corresponding to each sample feature vector to label the sample explosion-proof performance score label set; based on machine learning, constructing an explosion-proof performance evaluation model; using the sample feature vector set as input and the sample explosion-proof performance score label set as supervision signal, supervising the training of the explosion-proof performance evaluation model, and completing the training after the error convergence of the model output and the score label.

[0060] In the embodiments of the present application, first, a sample feature vector set is extracted from the historical operation multi-source sensor data of the explosion-proof diesel engine. The standardized sensor data sequence is segmented according to a preset sliding time window, the time domain statistical features are calculated, the frequency domain features are extracted by frequency domain transformation, and then the adaptive weighted fusion is performed according to the historical variation degree. At the same time, the true explosion-proof performance score corresponding to each sample feature vector is collected, which can be obtained by actual test, and after labeling, the sample explosion-proof performance score label set is formed.

[0061] Secondly, an explosion-proof performance evaluation model is constructed based on machine learning method. Specifically, a neural network is selected as the basic framework to build the explosion-proof performance evaluation model, and a multi-layer perceptron is constructed, including an input layer, a hidden layer and an output layer. The number of neurons in the input layer is determined according to the dimension of the sample feature vector, and the number of neurons in the output layer is 1, which is used to output the explosion-proof performance score value.

[0062] Then, the sample feature vector set is used as the input of the model, and the sample explosion-proof performance score label set is used as the supervision signal to supervise the training of the explosion-proof performance evaluation model. During the training process, the model will continuously adjust its parameters, so that the error between the output of the model and the score label gradually decreases. The loss function such as mean square error can be used to measure the difference between the output of the model and the score label, and the parameters of the model are updated through the back propagation algorithm. When the error converges to a small value, it is considered that the model has learned the mapping relationship between the feature vector and the explosion-proof performance score, and at this time the training of the model is completed.

[0063] Exemplarily, the explosion-proof performance evaluation model is constructed and trained based on neural network, and the specific steps are as follows: First, data acquisition, collecting the sample feature vector set of the historical operation multi-source sensor data of the explosion-proof diesel engine and the true value of the explosion-proof performance score corresponding to each sample feature vector, and constructing the explosion-proof performance score label set. The input nodes of the explosion-proof performance evaluation model are the sample explosion-proof performance score label set.

[0064] Secondly, model construction, a convolutional neural network model is constructed, including convolutional layer, pooling layer, fully connected layer, etc. The convolutional layer is used to extract the features of the data, the pooling layer is used to reduce the size of the feature data, and the fully connected layer is used to convert the feature data into a label for evaluating the explosion-proof performance. The input layer has a node number equal to the dimension of the input feature, for example, if the sample feature vector set has 10 features, the input layer contains 10 nodes; 1-3 hidden layers are set, the number of nodes in each layer is adjusted through experiments, such as 64, 32, etc., and the activation function is selected as ReLU; the number of nodes of the output layer is equal to the performance requirement of the evaluation, such as outputting an explosion-proof performance score value consuming 1 node, and the output layer generally does not use an activation function, and directly outputs a continuous value.

[0065] Thirdly, model training, the explosion-proof performance score value of the evaluation is used as the output. The sample explosion-proof performance score label set is used as a supervision signal, an Adam optimizer and a mean square error loss function are used to construct a training framework, the batch size is set to 32, the total training rounds are set to 50, and a patience mechanism (patience=5) is introduced, when the validation set loss does not appear for 5 consecutive rounds, the training process is automatically terminated, and the trained explosion-proof performance evaluation model is obtained, which can effectively avoid model overfitting and ensure that the model reaches a convergence state. Multiple configuration unit architectures are trained synchronously to obtain multiple pulse width configuration units.

[0066] Finally, based on the built and trained explosion-proof performance evaluation model, the temperature, pressure and combustible gas concentration of the explosion-proof diesel engine during the running process collected by the multi-source sensor are input, and the explosion-proof performance score value is output through the explosion-proof performance evaluation model. The explosion-proof performance score value can directly reflect the pros and cons of the explosion-proof performance of the explosion-proof diesel engine under the current running state. If the score value is high, it means that the explosion-proof performance of the explosion-proof diesel engine under the current temperature, pressure and combustible gas concentration conditions is good, and the explosion-proof diesel engine can be reliably operated, and the current running mode and state can be maintained to reduce the frequency of inspection and maintenance.

[0067] S30: If the explosion-proof performance score value is lower than the preset safety threshold, an operation parameter adjustment strategy is generated to dynamically adjust the operation parameters of the diesel engine; In the embodiment of the present application, when the explosion-proof performance score value output by the explosion-proof performance evaluation model is lower than the preset safety threshold, it means that the current running state of the explosion-proof diesel engine has safety hazards, and timely measures need to be taken for adjustment. Based on the analysis of multi-source sensor data and the learning of the relationship between the model and the explosion-proof performance of each parameter, an operation parameter adjustment strategy is generated.

[0068] First, according to the previously fused feature vector and the correlation degree between each feature and the explosion-proof performance learned during model training, the abnormal operation parameters causing the decline of explosion-proof performance are determined. For example, if the weight of the combustible gas concentration feature vector is large and the current concentration exceeds the normal range, then the emission-related parameters of the combustible gas may be the focus of adjustment.

[0069] Exemplarily, the characteristics of engine operation in the target explosion-proof diesel engine are as shown in Figure 4 where the increase of carbon monoxide beyond the normal range poses a safety hazard.

[0070] Then, based on the analysis results, specific operation parameter adjustment strategies are generated. Different adjustment methods are adopted for different key factors. For example, if the temperature is too high, causing poor explosion-proof performance, the operation parameters of the cooling system can be adjusted, such as increasing the flow of coolant or increasing the speed of the cooling fan, to reduce the working temperature of the diesel engine; if the pressure is unstable, the valve opening of the intake or exhaust system may need to be adjusted to balance the pressure fluctuation.

[0071] Finally, after determining the adjustment strategy, the operation parameters of the diesel engine are dynamically adjusted. Each adjustment is of a small amplitude, and the change of the explosion-proof performance score value is monitored in real time. If the score value improves, it means that the adjustment strategy is effective, and the same direction can be continued for fine-tuning until the score value reaches or exceeds the preset safety threshold. If the score value does not improve or even further decreases, the cause is reanalyzed, the strategy is adjusted, and the attempt is made again.

[0072] Specifically, the operation parameter adjustment strategy is generated, including: calculating a basic adjustment amplitude of the operation parameter according to the deviation degree of the explosion-proof performance score value from the preset safety threshold; obtaining real-time load parameters and external environment temperature data of the explosion-proof diesel engine; dynamically correcting the basic adjustment amplitude in combination with the real-time load parameters and external environment temperature data; configuring a dynamic adjustment coefficient based on the frequency energy accumulation rate; multiplying the corrected basic adjustment amplitude by the dynamic adjustment coefficient to obtain the final adjustment amount of fuel injection quantity, intake quantity, and cooling intensity.

[0073] In one application embodiment, first, a basic adjustment amplitude of the operation parameter is calculated according to the deviation degree of the explosion-proof performance score value from the preset safety threshold. The greater the deviation, the greater the basic adjustment amplitude. For example, when the deviation reaches a certain degree, the basic adjustment amplitude can be appropriately increased to quickly make the explosion-proof performance score value close to or exceed the preset safety threshold.

[0074] Exemplarily, the full score is 100 points, and the higher the score, the better the explosion-proof performance. Assuming that the preset safety threshold is 80 points: If the explosion-proof performance score is 75 points, it is lower than the threshold, and the deviation is 5 points from the threshold, the deviation rate = 5 / 80 = 6.25%, and the setting deviation rate corresponds to a basic adjustment amplitude of 1% per 1%. Therefore, the basic adjustment amplitude of the fuel injection amount is -6.25%, the injection amount is reduced to reduce the combustion temperature; the basic adjustment amplitude of the intake amount is +6.25%, the intake amount is increased to dilute the combustible mixture; and the basic adjustment amplitude of the cooling intensity is +6.25%, the cooling is enhanced to reduce the temperature of the engine.

[0075] If the explosion-proof performance score is 85 points, it is higher than the threshold, and the deviation is -5 points, the deviation rate = -5 / 80 = -6.25%, and the setting deviation rate corresponds to a basic adjustment amplitude of 0.5% per 1%. Therefore, the basic adjustment amplitude of the fuel injection amount is +3.125%, the injection amount is slightly increased to ensure power; the intake amount is -3.125%, and the cooling intensity is -3.125%.

[0076] Secondly, real-time load parameters and external environment temperature data of the explosion-proof diesel engine are obtained. The real-time load parameters reflect the current working load of the diesel engine, and the external environment temperature will affect the operation of the diesel engine. For example, in a high temperature environment, the heat dissipation difficulty of the diesel engine will increase, and the cooling intensity needs to be adjusted more significantly; and when running under high load, the fuel injection amount and the intake amount may need to be increased accordingly.

[0077] Then, the basic adjustment amplitude is dynamically corrected by combining the real-time load parameters and the external environment temperature data. That is, after calculating the basic adjustment amplitude of the explosion-proof diesel engine operating parameters, the basic adjustment amplitude needs to be adjusted according to the actual working load and the actual environment temperature of the diesel engine, so as to be more suitable for the current operating conditions.

[0078] Further, if the real-time load is large and the external environment temperature is high, the adjustment amount needs to be appropriately increased based on the basic adjustment amplitude; on the contrary, if the load is small and the environment temperature is suitable, the adjustment amount can be appropriately reduced.

[0079] Exemplarily, assuming that the basic adjustment amplitude of the fuel injection amount is -6.25%, the basic adjustment amplitude of the intake amount is +6.25%, and the basic adjustment amplitude of the cooling intensity is +6.25%. The load-environment temperature-adjustment coefficient mapping table is queried in combination with the real-time working condition of the explosion-proof diesel engine: if the current diesel engine load is 75%, it is in the medium-high load interval; the external environment temperature is 32°C, which is in the medium-high temperature interval, and the corresponding correction coefficient can be obtained by looking up the table.

[0080] Further, the fuel injection amount correction coefficient 1.1, the intake amount correction coefficient 1.05, and the cooling intensity correction coefficient 1.15 are obtained by looking up the table.

[0081] The base adjustment range is multiplied by the corresponding correction coefficient to obtain the final adjustment amount: the final adjustment amount of fuel injection quantity = -6.25% x 1.1 = -6.875%, and the actual fuel injection quantity is reduced by 6.875%.

[0082] The final adjustment amount of intake air quantity = +6.25% x 1.05 = +6.5625%, and the actual intake air quantity is increased by 6.5625%.

[0083] The final adjustment amount of cooling intensity = +6.25% x 1.15 = +7.1875%, and the actual cooling intensity is increased by 7.1875%.

[0084] Again, based on the frequency energy accumulation rate, a dynamic adjustment coefficient is configured. The frequency energy accumulation rate reflects the changes in the system operating state. When the frequency energy accumulation rate is fast, it means that the system fluctuates greatly, and more aggressive adjustment may be needed. At this time, the dynamic adjustment coefficient can be set larger. When the frequency energy accumulation rate is slow, it means that the system is relatively stable, and the dynamic adjustment coefficient can be set smaller.

[0085] Further, the dynamic adjustment coefficient is configured, that is, according to the time when the frequency energy accumulates from 0 to the total energy preset proportion, such as 80%, as the frequency energy accumulation rate index. The shorter the time, the faster the energy accumulates, and the greater the rate. A greater rate establishes a corresponding relationship with a greater adjustment coefficient. Finally, according to the actual calculated rate index, the corresponding relationship is found to obtain the fitted dynamic adjustment coefficient.

[0086] For example, in the explosion-proof diesel engine test, when calculating the dominant frequency component of the exhaust port pressure, if the frequency energy accumulates from zero to 80% of the total energy only in 8 seconds, the accumulation rate is large, indicating that the key fluctuation frequency of the pressure changes quickly and the operating state is unstable. According to the positive mapping relationship established in advance, 8 seconds correspond to an adjustment coefficient of 1.5, and 1.5 is taken as the dynamic adjustment coefficient. If it takes 25 seconds to accumulate to 80% of the total energy, the accumulation rate is small, the key fluctuation frequency of the pressure changes smoothly, and the operating state is stable. The mapping relationship shows that 25 seconds correspond to an adjustment coefficient of 0.9, and 0.9 is taken as the dynamic adjustment coefficient.

[0087] Finally, the corrected base adjustment range is multiplied by the dynamic adjustment coefficient to obtain the final adjustment amount of fuel injection quantity, intake air quantity, and cooling intensity. For example, the corrected base adjustment range determines the preliminary adjustment range of fuel injection quantity, intake air quantity, and cooling intensity, and then multiplied by the dynamic adjustment coefficient to obtain more accurate final adjustment amount.

[0088] The above adjustment strategy ensures that the adjustment of the diesel engine operating parameters is more scientific and reasonable, effectively improves the explosion-proof performance, makes the explosion-proof performance score value reach or exceed the preset safety threshold as soon as possible, and ensures the safe and stable operation of the explosion-proof diesel engine.

[0089] Further, based on the frequency domain energy accumulation rate, a dynamic adjustment coefficient is configured, including: extracting the time required for the frequency domain energy to accumulate from zero to reach the total energy preset percentage threshold in the process of calculating the dominant frequency component as the frequency domain energy accumulation rate index; establishing a positive mapping relationship between the frequency domain energy accumulation rate and the adjustment coefficient, wherein the greater the frequency domain energy accumulation rate value, the greater the corresponding adjustment coefficient; According to the frequency domain energy accumulation rate index, the positive mapping relationship is queried to obtain the corresponding dynamic adjustment coefficient.

[0090] In the embodiments of the present application, first, the time required for the frequency domain energy to accumulate from zero to reach the total energy preset percentage threshold in the process of calculating the dominant frequency component is extracted as the frequency domain energy accumulation rate index. Reflect the change of system running state, if the time is short, it means that the frequency domain energy accumulation rate is fast, and the system fluctuates greatly; if the time is long, it means that the frequency domain energy accumulation rate is slow, and the system is relatively stable.

[0091] Secondly, a positive mapping relationship between the frequency domain energy accumulation rate and the adjustment coefficient is established. Since the greater the frequency domain energy accumulation rate value, the greater the system fluctuation, more aggressive adjustment is needed, so the corresponding adjustment coefficient is also greater; on the contrary, the smaller the frequency domain energy accumulation rate value, the more stable the system, the smaller the adjustment coefficient.

[0092] Exemplarily, through the test of the explosion-proof diesel engine under different working conditions, the actual value range of the rate index is counted, such as 5 seconds to 30 seconds. Then, the rate interval is divided into 3-5 gradient intervals according to the time from short to long, i.e. the rate from large to small, and the initial adjustment coefficient is set for each interval.

[0093] For example, 5 seconds to 10 seconds corresponds to 1.3-1.5, 11 seconds to 20 seconds corresponds to 1.0-1.2, and 21 seconds to 30 seconds corresponds to 0.8-0.9.

[0094] Finally, according to the frequency domain energy accumulation rate index, the positive mapping relationship is queried to obtain the corresponding dynamic adjustment coefficient. Multiply the dynamic adjustment coefficient with the corrected basic adjustment amplitude to obtain the final adjustment amount of the more accurate fuel injection amount, intake amount and cooling intensity, so as to realize the scientific and reasonable adjustment of the running parameters of the explosion-proof diesel engine, further improve the explosion-proof performance, and ensure that the explosion-proof diesel engine is always in a safe and stable running state.

[0095] S40: Re-collect the adjusted multi-source sensing data, verify whether the explosion-proof performance score value reaches the preset safety threshold, and if it reaches, determine that the explosion-proof performance is qualified.

[0096] In the embodiments of this application, after dynamically adjusting the operating parameters of the diesel engine, the multi-source sensing data after adjustment is collected again, including data such as temperature, pressure, and combustible gas concentration. According to the previous feature extraction method, time-domain statistical features and frequency-domain features are extracted from the newly collected data, and then adaptive weighted fusion is performed to form a new feature vector. The new feature vector is input into the trained explosion-proof performance evaluation model to obtain a new explosion-proof performance score value.

[0097] Then, the new explosion-proof performance score value is compared with the preset safety threshold. If the score value reaches or exceeds the preset safety threshold, it indicates that the explosion-proof performance of the explosion-proof diesel engine has been effectively improved by adjusting the operating parameters. At this time, it is determined that the explosion-proof performance is qualified, and the diesel engine can continue to operate safely.

[0098] However, if the new score value is still lower than the preset safety threshold, it means that the previous adjustment strategy may not have achieved the expected effect, or there are factors affecting the explosion-proof performance that have not been discovered. It is necessary to analyze the multi-source sensing data again, re-determine the abnormal operating parameters, generate a new operating parameter adjustment strategy, dynamically adjust the operating parameters of the diesel engine again, and then repeat the process of collecting data and evaluating performance until the explosion-proof performance score value reaches the preset safety threshold and the explosion-proof performance is determined to be qualified.

[0099] In summary, compared with the prior art, this application constructs an explosion-proof performance evaluation model by adaptively weighted fusing the features of multi-source sensing data, and accurately evaluates the explosion-proof performance of the explosion-proof diesel engine. At the same time, according to the evaluation results, the operating parameters are dynamically adjusted to eliminate potential safety hazards in a timely manner and ensure that the explosion-proof diesel engine operates in a safe state. By continuously optimizing the adjustment strategy, the accuracy and reliability of the explosion-proof performance evaluation are improved to meet the requirements of different working conditions and environments.

[0100] In summary, the embodiments of this application at least have the following technical effects: In the embodiments of this application, by providing a method and device for testing the explosion-proof performance of an explosion-proof diesel engine system based on intelligent sensing, first, the explosion-proof performance of the explosion-proof diesel engine system is evaluated by using the multi-source sensing data during the operation of the explosion-proof diesel engine. Secondly, based on the multi-source sensing data, a scientific explosion-proof performance evaluation model is constructed to avoid the limitations of only focusing on the determination of a single parameter threshold and improve the accuracy of the explosion-proof performance evaluation. At the same time, a dynamic operating parameter adjustment strategy is generated based on the real-time multi-source sensing data, overcoming the problem that parameter adjustment depends on an empirical fixed strategy and enhancing the adaptability of the system to different working conditions. Through the above technical solutions, it is possible to accurately evaluate and effectively adjust the explosion-proof performance of the explosion-proof diesel engine system, improve the test efficiency and reliability. The in-depth excavation and comprehensive utilization of multi-source sensing data can timely discover potential safety hazards, provide an effective decision-making basis for production and maintenance, and reduce the risk of safety accidents.

[0101] Embodiment two, as shown in Figure 2 the same inventive concept as the explosion-proof performance test method for the intelligent sensing-based explosion-proof diesel engine system provided in embodiment one, the embodiment of the present application also provides an explosion-proof performance test device for the intelligent sensing-based explosion-proof diesel engine system, comprising: A data acquisition module 11 is configured to acquire multi-source sensing data of the explosion-proof diesel engine during operation, wherein the multi-source sensing data at least includes temperature, pressure and combustible gas concentration. A model construction module 12 is configured to construct an explosion-proof performance evaluation model based on the multi-source sensing data, and output an explosion-proof performance score value. A parameter adjustment module 13 is configured to generate an operation parameter adjustment strategy to dynamically adjust the operation parameters of the diesel engine if the explosion-proof performance score value is lower than a preset safety threshold. A performance determination module 14 is configured to reacquire the adjusted multi-source sensing data, verify whether the explosion-proof performance score value reaches the preset safety threshold, and determine that the explosion-proof performance is qualified if it does.

[0102] In one embodiment, the data acquisition module 11 is specifically configured to: acquire temperature, pressure and combustible gas concentration data in real time through intelligent sensor nodes arranged at the exhaust port, the air inlet and the engine cabin of the explosion-proof diesel engine; perform time sequence alignment and noise filtering processing on the acquired data to form a standardized sensing data sequence.

[0103] In one embodiment, the model construction module 12 is specifically configured to: extract a feature vector of the multi-source sensing data; input the feature vector into a pre-trained explosion-proof performance evaluation model to output an obtained explosion-proof performance score value, wherein the higher the explosion-proof performance score value is, the better the explosion-proof performance is.

[0104] wherein the feature vector of the multi-source sensing data is extracted by: segmenting the standardized sensing data sequence according to a preset sliding time window; calculating corresponding time domain statistical features including maximum value, minimum value, mean value and standard deviation for the temperature, pressure and combustible gas concentration data in each time window; performing frequency domain transformation on the data in each time window, and extracting the amplitude of the dominant frequency component as a frequency domain feature; combining the time domain statistical features and the frequency domain features, and adaptively weighting and fusing based on the historical variation degree of each sensing data to form a final feature vector.

[0105] Further, in one application embodiment, the data in each time window is subjected to frequency domain transformation, and the amplitude of the dominant frequency component is extracted as a frequency domain feature, including: Performing fast Fourier transform on the sensor data sequence in each time window, converting the sensor data sequence from time domain to frequency domain, and obtaining the corresponding frequency domain amplitude sequence; Calculating the total energy of all frequency components in the frequency domain amplitude sequence; Sorting the frequency components according to the amplitude, and accumulating the energy from the highest amplitude component until the accumulated energy reaches the preset percentage threshold of the total energy; The frequency component corresponding to the energy involved in the accumulation is determined as the dominant frequency component, and the corresponding amplitude is recorded as the frequency domain feature.

[0106] Further, in one application embodiment, the time domain statistical features and the frequency domain features are combined, and an adaptive weighted fusion is performed based on the historical variation degree of each sensor data to form a final feature vector, including: The time domain statistical features and the frequency domain features extracted from the temperature, pressure and combustible gas concentration data are spliced in the order of sensor type to form an initial fusion feature vector; Calculating the coefficient of variation of each sensor data under the historical normal operating state, wherein the coefficient of variation is the ratio of the standard deviation to the mean; Taking the reciprocal of the coefficient of variation as the basis value of each feature weight, and obtaining the adaptive weight coefficient corresponding to each feature after normalization processing; Constructing a diagonal weight matrix with the adaptive weight coefficient as the diagonal element, multiplying the initial fusion feature vector by the diagonal weight matrix to obtain the weighted final feature vector.

[0107] Further, in one application embodiment, the training step of the explosion-proof performance evaluation model includes: Extracting a sample feature vector set of the historical running multi-source sensor data of the explosion-proof diesel engine, and collecting the explosion-proof performance score true value corresponding to each sample feature vector to obtain a sample explosion-proof performance score label set; Based on machine learning, constructing an explosion-proof performance evaluation model; Using the sample feature vector set as input and the sample explosion-proof performance score label set as supervision signal, the explosion-proof performance evaluation model is supervised trained, and the training is completed after the error convergence of the model output and the score label.

[0108] Further, in one application embodiment, a dynamic adjustment coefficient is configured based on the frequency domain energy accumulation rate, including: Extract the time required for the frequency domain energy to accumulate from zero to a preset percentage threshold of the total energy during the calculation of the dominant frequency component, and use this as the frequency domain energy accumulation rate index. A positive mapping relationship is established between the frequency domain energy accumulation rate and the adjustment coefficient, wherein the larger the frequency domain energy accumulation rate value, the larger the corresponding adjustment coefficient; Based on the frequency domain energy accumulation rate index, the positive mapping relationship is queried to obtain the corresponding dynamic adjustment coefficient.

[0109] When the operating speed of an explosion-proof diesel engine changes, its operating characteristics also change, such as... Figure 4 As shown: As the operating speed of an explosion-proof diesel engine increases, its power output also increases, along with exhaust temperature and pressure. The high-temperature, high-pressure exhaust environment places higher demands on its explosion-proof performance, as high temperatures may cause combustible gases to reach their ignition point, while high pressure may lead to dangerous situations such as gas leaks. Simultaneously, increased engine speed intensifies friction between components, generating more heat and increasing the system's thermal load. Further optimization of the cooling system is necessary to ensure that the cooling intensity meets the system's heat dissipation requirements and prevent safety issues caused by overheating.

[0110] While lower operating speeds reduce power output and lower system heat load and pressure, incomplete combustion may occur. Incomplete combustion leads to an increase in the content of combustible gases in the exhaust, which also affects explosion-proof performance. Furthermore, system stability may be affected at low speeds, and the operating status of various components may become unstable, thus impacting the accuracy and reliability of multi-source sensor data.

[0111] At different engine speeds, parameters such as temperature, pressure, and combustible gas concentration in multi-source sensor data will change significantly. For example, as the engine speed increases, the temperature sensor may detect a significant increase in temperature at the exhaust port and inside the engine compartment; the pressure sensor will record pressure fluctuations and increases; and the combustible gas concentration sensor may detect fluctuations in the content of combustible gas within a certain range.

[0112] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0113] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0114] The specification and drawings are, of course, to be regarded in an illustrative rather than a restrictive sense. It is to be understood that any such modifications, variations, combinations or equivalents that fall within the scope of the application are intended to be embraced herein.

Claims

1. A method for testing the explosion-proof performance of an intelligent sensor-based explosion-proof diesel engine system, characterized in that, The method comprises: acquiring multi-source sensing data of the explosion-proof diesel engine during operation, wherein the multi-source sensing data at least includes temperature, pressure and combustible gas concentration; based on the multi-source sensing data, constructing an explosion-proof performance evaluation model to output an explosion-proof performance score value; if the explosion-proof performance score value is lower than a preset safety threshold, generating an operation parameter adjustment strategy to dynamically adjust the operation parameters of the diesel engine; recollecting the adjusted multi-source sensing data to verify whether the explosion-proof performance score value reaches the preset safety threshold, and if so, determining that the explosion-proof performance is qualified.

2. The method according to claim 1, wherein Acquiring multi-source sensing data of the explosion-proof diesel engine during operation comprises: real-time collecting temperature, pressure and combustible gas concentration data through intelligent sensor nodes arranged at the exhaust port, the air inlet and the engine cabin of the explosion-proof diesel engine; aligning the collected data in time sequence and performing noise filtering processing to form a standardized sensing data sequence.

3. The method according to claim 1, wherein Based on the multi-source sensing data, constructing an explosion-proof performance evaluation model to output an explosion-proof performance score value comprises: extracting a feature vector of the multi-source sensing data; inputting the feature vector into a pre-trained explosion-proof performance evaluation model to output an obtained explosion-proof performance score value, wherein the higher the explosion-proof performance score value is, the better the explosion-proof performance is.

4. The method according to claim 3, wherein, Extracting a feature vector of the multi-source sensing data comprises: segmenting the standardized sensing data sequence according to a preset sliding time window; for the temperature, pressure and combustible gas concentration data in each time window, respectively calculating corresponding time domain statistical features including maximum value, minimum value, mean value and standard deviation; performing frequency domain transformation on the data in each time window to extract the amplitude of the dominant frequency component as a frequency domain feature; combining the time domain statistical features and the frequency domain features, and adaptively weighting and fusing based on the historical variation degree of each sensing data to form a final feature vector.

5. The method according to claim 4, wherein the method is used for testing the explosion-proof performance of the intelligent-sensor-based explosion-proof diesel engine system. Performing frequency domain transformation on the data in each time window to extract the amplitude of the dominant frequency component as a frequency domain feature comprises: performing fast Fourier transform on the sensing data sequence in each time window to convert the sensing data sequence from time domain to frequency domain to obtain a corresponding frequency domain amplitude sequence; calculating the total energy of all frequency components in the frequency domain amplitude sequence; sorting the frequency components according to the amplitude size, and starting from the highest amplitude component to accumulate energy until the accumulated energy reaches a preset percentage threshold of the total energy; determining the frequency component corresponding to the accumulated energy as the dominant frequency component, and recording the corresponding amplitude as the frequency domain feature.

6. The method according to claim 5, wherein Combining the time domain statistical features and the frequency domain features, and adaptively weighting and fusing based on the historical variation degree of each sensing data to form a final feature vector comprises: splicing the time domain statistical features and the frequency domain features extracted from the temperature, pressure and combustible gas concentration data in the order of sensor types to form an initial fusion feature vector; calculating the coefficient of variation of each sensing data under the historical normal operation state, wherein the coefficient of variation is the ratio of the standard deviation to the mean value; taking the reciprocal of the coefficient of variation as the basis value of each feature weight, and obtaining the adaptive weight coefficient corresponding to each feature through normalization processing; A diagonal weight matrix with the adaptive weight coefficients as diagonal elements is constructed, and the initial fusion feature vector is multiplied by the diagonal weight matrix to obtain a weighted final feature vector.

7. The method according to claim 3, wherein The training step of the explosion-proof performance evaluation model comprises: A sample feature vector set of the historical operation multi-source sensing data of the explosion-proof diesel engine is extracted, and a corresponding explosion-proof performance score true value of each sample feature vector is collected to obtain a sample explosion-proof performance score label set; An explosion-proof performance evaluation model is constructed based on machine learning; The sample feature vector set is used as input, and the sample explosion-proof performance score label set is used as a supervision signal to supervise the training of the explosion-proof performance evaluation model, and the training is completed after the error convergence of the model output and the score label.

8. The method of testing the explosion-proof performance of an explosion-proof diesel engine system based on intelligent sensing according to claim 1, characterized in that, An operation parameter adjustment strategy is generated, comprising: According to the deviation degree of the explosion-proof performance score value and the preset safety threshold, the basic adjustment amplitude of the operation parameter is calculated; Real-time load parameters and external environment temperature data of the explosion-proof diesel engine are obtained; The real-time load parameters and external environment temperature data are combined to dynamically correct the basic adjustment amplitude; A dynamic adjustment coefficient is configured based on the frequency energy accumulation rate; The corrected basic adjustment amplitude is multiplied by the dynamic adjustment coefficient to obtain the final adjustment amount of the fuel injection amount, the intake amount and the cooling intensity.

9. The method according to claim 8, wherein the method is a method for testing the explosion-proof performance of an explosion-proof diesel engine system based on intelligent sensing, characterized in that, Configuring a dynamic adjustment coefficient based on the frequency energy accumulation rate comprises: Extracting the time required for the frequency energy to accumulate from zero to reach a preset percentage threshold of the total energy during the calculation of the dominant frequency component as a frequency energy accumulation rate indicator; A positive mapping relationship between the frequency energy accumulation rate and the adjustment coefficient is established, wherein the greater the frequency energy accumulation rate value, the greater the corresponding adjustment coefficient; According to the frequency energy accumulation rate indicator, the corresponding dynamic adjustment coefficient is obtained by querying the positive mapping relationship.

10. An intelligent sensor-based explosion-proof performance test device for an explosion-proof diesel engine system, characterized in that, The device for performing the method of any one of claims 1-9 comprises: A data acquisition module for acquiring multi-source sensing data of the explosion-proof diesel engine during operation, wherein the multi-source sensing data at least includes temperature, pressure and combustible gas concentration; A model construction module for constructing an explosion-proof performance evaluation model based on the multi-source sensing data, and outputting an explosion-proof performance score value; A parameter adjustment module for generating an operation parameter adjustment strategy and dynamically adjusting the operation parameters of the diesel engine if the explosion-proof performance score value is lower than a preset safety threshold; A performance determination module for re-collecting the adjusted multi-source sensing data to verify whether the explosion-proof performance score value reaches the preset safety threshold, and determining that the explosion-proof performance is qualified if it does.

Images