Frame performance detection method of electric sanitation tricycle

Abstract

The invention discloses a frame performance detection method for an electric sanitation tricycle, and relates to the related field of material detection, and the method comprises the steps: obtaining basic design information; searching is carried out, a plurality of welding abnormal positions and a plurality of welding abnormal records are obtained, and welding spot abnormity degree identification is carried out; welding spot abnormity identification is carried out on welding spots in the frame welding spot topology network; k central welding spots are randomly extracted from the vehicle frame welding spot abnormity degree identification topology network, neighborhood construction is carried out, and K vehicle frame welding spot abnormity degree identification sub-topology networks are obtained; carrying out welding spot anomaly weighted calculation to obtain welding spot anomaly weighted values of the K sub-topology networks; and determining the performance detection quantity of the K welding spots, and carrying out sampling performance detection. The technical problems that existing electric sanitation tricycle frame performance detection is not high in pertinence and poor in detection effect, and potential hidden dangers of a frame are difficult to find accurately are solved, and the technical effects of pertinence evaluation of the frame performance, accurate positioning of potential welding anomalies and detection accuracy improvement are achieved.

Description

Technical Field

[0001] The present application relates to the field of material testing, and in particular to a method for testing the frame performance of an electric three-wheeled sanitation vehicle. Background Art

[0002] As an important tool for urban sanitation work, the frame performance of electric three-wheeled sanitation vehicles is directly related to the durability, stability and safety of the vehicle. With the widespread use of electric three-wheeled sanitation vehicles, the demand for frame performance testing and evaluation is also increasing. Traditional frame performance testing methods often lack in-depth analysis of specific frame problems, resulting in wide coverage but poor results, and difficulty in accurately locating potential problems.

[0003] Among the current related technologies, the frame performance testing of electric three-wheeled sanitation vehicles is not very targeted, resulting in poor testing results and difficulty in accurately discovering technical problems with potential hidden dangers in the frame. Summary of the invention

[0004] The present application provides a frame performance detection method for an electric three-wheeled sanitation vehicle. The method obtains basic design information of the target electric three-wheeled sanitation vehicle, retrieves the historical maintenance database to obtain welding abnormality records and calculates the degree of abnormality of the weld point, constructs a frame weld point abnormality identification topological network, randomly extracts and authenticates the central weld point, constructs a sub-topological network, calculates the weighted value of the sub-topological network weld point abnormality, determines the detection amount according to the weighted value and performs sampling welding performance detection, and other technical means to achieve the technical effect of targeted evaluation of frame performance, accurately locates potential welding abnormalities, and improves detection accuracy.

[0005] The present application provides a method for testing the frame performance of an electric three-wheeled sanitation vehicle, including: The basic design information of the target electric three-wheeled sanitation vehicle is obtained, wherein the basic design information includes the sanitation vehicle model, frame material and frame welding point topological network; the historical electric three-wheeled sanitation vehicle maintenance database is searched with the sanitation vehicle model, frame material and welding abnormality maintenance as indexes to obtain multiple welding abnormality positions and multiple welding abnormality records, and the welding point abnormality degrees are identified for the multiple welding abnormality records to obtain multiple welding point abnormality degrees, wherein each welding abnormality position corresponds to a welding point abnormality degree; based on the multiple welding abnormality positions and the multiple welding point abnormality degrees, the welding points in the frame welding point topological network are marked with welding point abnormality degrees to obtain a frame welding point abnormality degree marking topological network; randomly select a welding point abnormality degree from the frame welding point abnormality degree marking topological network. Take K welds as K central welds, and perform topological network center point authentication on the K central welds; if the authentication is passed, construct neighborhoods in the frame weld abnormality identification topological network according to the K central welds and the corresponding K weld abnormality degrees, and obtain K frame weld abnormality identification sub-topological networks; traverse the K frame weld abnormality identification sub-topological networks to perform weld abnormality weighted calculations, and obtain K sub-topological network weld abnormality weighted values; determine K weld performance detection quantities according to the sizes of the K sub-topological network weld abnormality weighted values ​​and the preset welding performance detection total amount, and perform sampling welding performance detection on the K frame weld abnormality identification sub-topological networks based on the K weld performance detection quantities to obtain welding performance detection results.

[0006] In a possible implementation, K welding points are randomly selected from the frame welding point abnormality identification topological network as K central welding points, the K central welding points are authenticated by the topological network center point, and the following processing is performed: The K center welds are enumerated in pairs to obtain multiple center weld combinations; in combination with the frame weld abnormality identification topological network, distance calculations are performed on the multiple center weld combinations to obtain multiple center weld distances; and it is determined whether there is a center weld distance greater than a preset center weld distance threshold among the multiple center weld distances. If so, the authentication result is failed.

[0007] In a possible implementation, the following processing is performed: If not, the number of K solder joint anomalies corresponding to the K central solder joints that exceed the preset solder joint anomaly is counted to obtain the anomaly amount; when the anomaly amount is greater than the preset anomaly amount threshold, the authentication result is passed.

[0008] In a possible implementation, neighborhood construction is performed in the frame weld abnormality identification topological network according to the K central welds and the corresponding K weld abnormality degrees, K frame weld abnormality identification sub-topological networks are obtained, and the following processing is performed: Taking the K central welds as the neighborhood centers, initial neighborhood construction is performed in the frame weld abnormality identification topological network based on a preset diffusion bandwidth to obtain K initial weld neighborhoods, wherein each initial weld neighborhood includes a plurality of welds whose distance to the central weld is less than or equal to the preset diffusion bandwidth; neighborhood diffusion identification is performed on the K initial weld neighborhoods according to the preset diffusion bandwidth to determine K target diffusion weld neighborhoods; neighborhood adjustment is performed on the K target diffusion weld neighborhoods based on the frame weld abnormality identification topological network to obtain K adjusted diffusion weld neighborhoods; and the K frame weld abnormality identification sub-topological networks are constructed according to the positions of each weld in the K adjusted diffusion weld neighborhoods.

[0009] In a possible implementation, the following processing is performed: The number of solder joints in the K initial solder joint neighborhoods is counted, and the statistical results are respectively compared with the volumes of the K initial solder joint neighborhoods to obtain K initial solder joint neighborhood densities; the K initial solder joint neighborhoods are diffused outwardly based on the preset diffusion bandwidth to obtain K first diffusion solder joint neighborhoods; the number of solder joints in the K first diffusion solder joint neighborhoods is counted, and the statistical results are respectively compared with the volumes of the K first diffusion solder joint neighborhoods to obtain K first diffusion solder joint neighborhood densities; it is determined whether the density of the K first diffusion solder joint neighborhoods is greater than or equal to the density of the K initial solder joint neighborhoods, and if so, the K first diffusion solder joint neighborhoods are diffused outwardly according to the preset diffusion bandwidth until a preset diffusion number or a preset diffusion stop condition is met, to obtain K target diffusion solder joint neighborhoods, wherein the preset diffusion stop condition is that the difference between the diffusion solder joint neighborhood density of the current diffusion solder joint neighborhood and the solder joint neighborhood density of the previous diffusion solder joint neighborhood is less than or equal to the preset difference.

[0010] In a possible implementation, the following processing is performed: The K target diffusion solder joint neighborhoods are removed from the frame solder joint abnormality identification topological network, and the remaining solder joints are used as multiple discrete solder joints; the overlapping solder joints in the K target diffusion solder joint neighborhoods are extracted to obtain multiple overlapping solder joints; the multiple discrete solder joints and the multiple overlapping solder joints are respectively added / retained to the nearest target diffusion solder joint neighborhood to obtain the K adjusted diffusion solder joint neighborhoods.

[0011] In a possible implementation, the K frame weld abnormality identification sub-topological networks are traversed to perform weld abnormality weighted calculation, and the weld abnormality weighted values ​​of the K sub-topological networks are obtained, and the following processing is performed: The distance from each welding point in the K frame welding point abnormality identification sub-topological network to the K central welding points is traversed and calculated to obtain K welding point distance sets, wherein each welding point corresponds to a welding point distance; the ratio of each welding point distance in the K welding point distance sets to the sum of multiple welding point distances in the corresponding welding point distance set is respectively calculated to obtain K weighted calculation coefficient sets; the welding point abnormality identification corresponding to each welding point in the K frame welding point abnormality identification sub-topological network is weightedly calculated according to the K weighted calculation coefficient sets to obtain the weighted values ​​of the welding point abnormality of the K sub-topological networks.

[0012] In a possible implementation, the following processing is performed: A plurality of sample welding anomaly records and a plurality of sample weld spot anomaly degrees are obtained as training data; the training data is used to perform supervised training on a framework constructed based on a BP neural network, and a mapping relationship between welding anomaly records and weld spot anomaly degrees is learned until the training converges, thereby obtaining a trained anomaly degree recognition network layer; the anomaly degree recognition network layer is used to perform weld spot anomaly degree recognition on the plurality of welding anomaly records, thereby obtaining the plurality of weld spot anomaly degrees.

[0013] The frame performance detection method of the electric three-wheeled sanitation vehicle proposed in the present application first obtains the basic design information of the target electric three-wheeled sanitation vehicle, wherein the basic design information includes the sanitation vehicle model, frame material and frame weld point topological network, and then retrieves the historical electric three-wheeled sanitation vehicle maintenance database with the sanitation vehicle model, frame material and abnormal welding maintenance as indexes to obtain multiple abnormal welding positions and multiple abnormal welding records, and performs weld point abnormality identification on the multiple abnormal welding records to obtain multiple weld point abnormality degrees, wherein each abnormal welding position corresponds to a weld point abnormality degree, and then performs weld point abnormality identification on the weld points in the frame weld point topological network based on the multiple abnormal welding positions and the multiple weld point abnormality degrees to obtain a frame weld point abnormality identification topological network, and then randomly extracts K weld points from the frame weld point abnormality identification topological network. As K central welds, the K central welds are authenticated as the central points of the topological network. If the authentication is passed, neighborhoods are constructed in the frame weld anomaly degree identification topological network according to the K central welds and the corresponding K weld anomaly degrees, and K frame weld anomaly degree identification sub-topological networks are obtained. Then, the K frame weld anomaly degree identification sub-topological networks are traversed to perform weighted calculation of weld anomaly degrees, and K sub-topological network weld anomaly degree weighted values ​​are obtained. Finally, according to the size of the K sub-topological network weld anomaly degree weighted values ​​and the preset welding performance detection total amount, the K weld performance detection quantities are determined. Based on the K weld performance detection quantities, the K frame weld anomaly degree identification sub-topological networks are sampled for welding performance detection, and the welding performance detection results are obtained, thereby achieving the technical effect of targeted evaluation of frame performance, accurately locating potential welding anomalies, and improving detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the method according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0015] Figure 1 A schematic flow chart of a method for detecting the frame performance of an electric three-wheeled sanitation vehicle provided in an embodiment of the present application.

[0016] Figure 2 A schematic diagram of the process of topological network center point authentication of K center welding points in the frame performance detection method of the electric three-wheeled sanitation vehicle provided in the embodiment of the present application. DETAILED DESCRIPTION

[0017] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0018] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0019] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0020] The present application embodiment provides a method for testing the frame performance of an electric three-wheeled sanitation vehicle, such as Figure 1 As shown, the method includes: Step S100, obtain the basic design information of the target electric three-wheeled sanitation vehicle, wherein the basic design information includes the sanitation vehicle model, frame material and frame weld point topology network. Specifically, the basic design information about the target electric three-wheeled sanitation vehicle is collected from design documents, data provided by the manufacturer or an existing database. This information includes but is not limited to the model of the sanitation vehicle (such as a certain brand XX model, used to distinguish the design differences between different models), the material of the frame (such as steel, aluminum alloy, etc.) and the frame weld point topology network (i.e., the connection relationship diagram of all welds in the frame).

[0021] Step S200, using the sanitation vehicle model, the frame material and abnormal welding maintenance as indexes to search the historical electric three-wheeled sanitation vehicle maintenance database, obtain multiple abnormal welding positions and multiple abnormal welding records, and identify the abnormal welding point degree of the multiple abnormal welding records to obtain multiple abnormal welding point degrees, wherein each abnormal welding position corresponds to a welding point abnormal degree. Specifically, using the sanitation vehicle model, frame material and abnormal welding maintenance as keywords, search for relevant records in the historical electric three-wheeled sanitation vehicle maintenance database. Further analyze the retrieved multiple abnormal welding positions (specific locations where welding points have problems or failures recorded in historical records) and multiple abnormal welding records to identify the abnormal welding point degree of each abnormal welding position, that is, the possibility or severity of problems with the welding point at that position. The abnormal welding point degree can be a probability value, a score or other form of measurement.

[0022] In a possible implementation, step S200 further includes step S210, obtaining multiple sample welding abnormality records and multiple sample welding point abnormality degrees as training data. Specifically, a certain number of welding abnormality records are screened out from the historical electric three-wheeled sanitation vehicle maintenance database, and these records contain information such as welding abnormality location, abnormality description, and repair measures. These welding abnormality records are evaluated using existing evaluation criteria, and a welding point abnormality degree is assigned to each record. These sample welding abnormality records and the corresponding sample welding point abnormality degrees are organized into a training data set for subsequent neural network training. Step S220, using the training data to supervise the framework constructed based on the BP neural network, learn the mapping relationship between welding abnormality records and welding point abnormality degrees, until the training converges, and obtain a trained abnormality degree recognition network layer. Specifically, a framework based on the BP (back propagation) neural network is designed, the input layer of which receives the characteristic information of the welding abnormality record (such as abnormal location, abnormal type, etc.), and the output layer outputs the welding point abnormality degree. The neural network is trained using the training data obtained in step S210. During the training process, the input layer receives sample welding abnormality records, and the output layer generates predicted weld abnormality. The error between the predicted value and the actual sample weld abnormality is calculated, and the weight and bias of the neural network are adjusted by the back propagation algorithm to reduce the error. The above process is repeated until the training error of the neural network reaches a preset convergence standard or the training reaches a preset number of iterations. Step S230, the weld abnormality is identified for the multiple welding abnormality records using the abnormality identification network layer to obtain the multiple weld abnormalities. Specifically, after the training is completed in step S220, a neural network model that can accurately identify the weld abnormality is obtained. The welding abnormality record to be identified is input into the input layer of the neural network, and the predicted weld abnormality is obtained through the forward propagation process of the network. These predicted weld abnormalities are used as inputs for subsequent steps to construct a frame weld abnormality identification topological network. This implementation adopts an abnormality recognition method based on BP neural network. BP neural network has powerful nonlinear mapping ability and self-learning ability. It can learn the complex mapping relationship between welding abnormality records and weld abnormality from a large amount of sample data. Through supervised training, the neural network can gradually approach the real mapping relationship, thereby realizing accurate recognition of new welding abnormality records and improving the accuracy and efficiency of weld abnormality recognition.

[0023] Step S300, based on the multiple welding abnormality positions and the multiple welding point abnormality degrees, the welding points in the frame welding point topological network are marked with welding point abnormality degrees, and the frame welding point abnormality degree marking topological network is obtained. Specifically, the obtained welding point abnormality degrees are combined with the frame welding point topological network, and the abnormality degree of each welding point in the topological network is marked. In this way, each welding point not only retains its position information in the frame, but also adds the attribute of abnormality. That is, the frame welding point abnormality degree marking topological network is a comprehensive network diagram that combines welding point positions, connection relationships and abnormality degree information.

[0024] Step S400, randomly extract K welding points from the frame welding point abnormality identification topological network as K central welding points, and perform topological network center point authentication on the K central welding points. Specifically, randomly extract K welding points from the frame welding point topological network with identified abnormality as central welding points, that is, welding points selected as analysis starting points in the frame welding point abnormality identification topological network, and authenticate these central welding points to ensure that they can serve as reliable starting points for subsequent analysis. The authentication process includes checking whether the distribution of the central welding points is uniform and whether they cover the key areas of the frame.

[0025] like Figure 2 As shown, in a possible implementation, K welding points are randomly selected from the frame welding point abnormality identification topological network as K central welding points, and the K central welding points are authenticated as the central points of the topological network. Step S400 further includes step S410, enumerating the K central welding points in pairs to obtain multiple central welding point combinations. Specifically, any two welding points are selected from the K central welding points randomly selected from the frame welding point abnormality identification topological network to form a central welding point combination. Repeat the above process until all possible central welding point combinations are enumerated. These combinations are used for subsequent distance calculation and authentication judgment. Step S420, combined with the frame welding point abnormality identification topological network, the distance calculation is performed on the multiple central welding point combinations respectively to obtain multiple central welding point distances. Specifically, for each central welding point combination, according to the welding point position information in the frame welding point abnormality identification topological network, the distance between the two welding points in the topological network is calculated, and this distance can be a physical distance (such as a straight-line distance) or a distance metric based on a topological structure. The distance calculation results of all center solder joint combinations are saved for subsequent certification judgment.

[0026] Step S430, it is determined whether there is a center welding point distance greater than the preset center welding point distance threshold value among the multiple center welding point distances. If so, the authentication result is not passed. Specifically, a preset center welding point distance threshold value is set, and this threshold value is determined according to the layout of the frame welding point topological network, the relative position relationship between the welding points, and the detection requirements. Traverse the calculation results of all center welding point distances to determine whether there is a distance greater than the preset center welding point distance threshold value. If there is a distance greater than the threshold value, it indicates that the distribution of the selected center welding point in the topological network is too dispersed, which is not conducive to the subsequent neighborhood construction and performance detection, so the authentication result is not passed. Step S440, if not, then count the number of K welding point abnormalities corresponding to the K center welding points that exceed the preset welding point abnormality, and obtain the abnormality amount. Specifically, if all center welding point distances are less than or equal to the preset center welding point distance threshold value, the welding point abnormality of the center welding point is further considered. Set a preset welding point abnormality degree to determine whether the degree of welding point abnormality is significant. Count the number of welding point abnormalities that exceed the preset welding point abnormality degree among the K center welding points, that is, the abnormality amount. Step S450, when the abnormal amount is greater than the preset abnormal amount threshold, the authentication result is passed. Specifically, a preset abnormal amount threshold is set, which can be determined according to factors such as the detection requirements and the overall abnormality of the frame welds, and is used to determine whether the central weld has a high representativeness. If the statistically obtained abnormal amount is greater than the preset abnormal amount threshold, it indicates that the selected central weld has a high representativeness and can better reflect the overall abnormality of the frame weld, so the authentication result is passed. This implementation method can determine whether the distribution of the central weld in the topological network is too dispersed by enumerating the central weld combinations and calculating the distances in pairs, thereby avoiding errors caused by uneven distribution of welds during subsequent performance testing. At the same time, by counting the number of weld abnormalities of the central weld that exceed the preset threshold, it can ensure that the selected central weld can better reflect the overall abnormality of the frame weld, and improve the accuracy and effectiveness of performance testing. This authentication method combines the position information and abnormality information of the weld, ensuring that the selected central weld is both reasonably distributed in the frame weld abnormality identification topological network and has a high abnormal representativeness.

[0027] Step S500, if the authentication is passed, neighborhoods are constructed in the frame weld abnormality identification topological network according to the K center welds and the corresponding K weld abnormality degrees, and K frame weld abnormality identification sub-topological networks are obtained. Specifically, starting from the authenticated center weld, neighborhoods are constructed in the frame weld abnormality identification topological network based on a preset analysis range. These neighborhoods form K frame weld abnormality identification sub-topological networks, each of which contains welds and their abnormality information within a certain range.

[0028] In a possible implementation, neighborhood construction is performed in the frame weld abnormality identification topological network according to the K center welds and the corresponding K weld abnormality degrees, and K frame weld abnormality identification sub-topological networks are obtained. Step S500 further includes step S510, taking the K center welds as the neighborhood centers, and performing initial neighborhood construction in the frame weld abnormality identification topological network based on a preset diffusion bandwidth to obtain K initial weld neighborhoods, wherein each initial weld neighborhood includes a plurality of welds whose distance to the center weld is less than or equal to the preset diffusion bandwidth. Specifically, the positions of the K center welds are determined, and a preset diffusion bandwidth is set, and this bandwidth determines the initial neighborhood range of each center weld. The size of the bandwidth can be determined according to the layout of the frame weld topological network, the relative position relationship between the welds, and the detection requirements. For each center weld, traverse the other welds in the frame weld abnormality identification topological network, calculate their distances to the center weld, and include the welds whose distance is less than or equal to the preset diffusion bandwidth into the initial neighborhood of the center weld to form K initial weld neighborhoods.

[0029] Step S520, performing neighborhood diffusion identification on the K initial solder joint neighborhoods according to the preset diffusion bandwidth, and determining K target diffusion solder joint neighborhoods. Specifically, diffusion identification is performed on each initial solder joint neighborhood, and whether the neighborhood range needs to be further expanded. The basis for diffusion identification can be the abnormality distribution of solder joints in the neighborhood, the connection relationship between solder joints, etc. If diffusion is identified as needed, the direction and range of diffusion are determined according to preset rules (such as based on the gradient change of solder joint abnormality, the connection tightness between solder joints, etc.), and the neighborhood range of each central solder joint is updated to form K target diffusion solder joint neighborhoods.

[0030] Step S530, based on the frame weld abnormality identification topological network, the neighborhood of the K target diffusion weld points is adjusted to obtain K adjusted diffusion weld point neighborhoods. Specifically, each target diffusion weld point neighborhood is traversed to check whether there are unreasonable weld points (such as weld points with extremely high or low abnormality, weld points with abnormal connection relationship with other weld points, etc.). If an unreasonable weld point is found, it is adjusted according to the information of the frame weld abnormality identification topological network, such as removing it from the neighborhood or replacing it with other more reasonable weld points. The neighborhood range of each central weld point is updated to form K adjusted diffusion weld point neighborhoods.

[0031] Step S540, constructing the K frame weld point abnormality identification sub-topological networks according to the positions of each weld point in the K adjusted diffusion weld point neighborhoods. Specifically, for each adjusted diffusion weld point neighborhood, the weld points and their position information are retained. A sub-topological network is constructed according to the connection relationship between the weld points (such as the connection relationship in the frame weld point topological network). In the sub-topological network, the abnormality information of the weld points is retained to form K frame weld point abnormality identification sub-topological networks. This implementation method performs initial neighborhood construction by presetting the diffusion bandwidth, ensuring that the neighborhood range covers the weld points that are relatively close to the central weld point. Then, through neighborhood diffusion identification and neighborhood adjustment, the neighborhood range is further optimized to make it more in line with the detection requirements, ensuring that the neighborhood range of each central weld point is both reasonable and can reflect the overall abnormality of the frame weld point. The frame weld point abnormality identification sub-topological network is constructed according to the adjusted neighborhood range, which provides a more reliable basis for subsequent performance testing and improves the accuracy and effectiveness of performance testing.

[0032] In a possible implementation, step S520 further includes step S521, counting the number of welds in the K initial weld neighborhoods, and respectively comparing the statistical results with the volumes of the K initial weld neighborhoods to obtain the K initial weld neighborhood densities. Specifically, for each initial weld neighborhood, the number of welds contained therein is calculated. Based on the spatial distribution of the weld neighborhood in the frame weld abnormality identification topological network, the volume of each initial weld neighborhood is determined. The number of welds in each initial weld neighborhood is divided by the volume of the neighborhood to obtain the initial weld neighborhood density.

[0033] Step S522, based on the preset diffusion bandwidth, the K initial solder point neighborhoods are diffused outward to obtain K first diffusion solder point neighborhoods. Specifically, for each initial solder point neighborhood, it is expanded outward according to the preset diffusion bandwidth to form a larger solder point neighborhood range, and the new solder point neighborhood (i.e., the first diffusion solder point neighborhood) includes all solder points in the initial solder point neighborhood and may include additional solder points. Step S523, the number of solder points in the K first diffusion solder point neighborhoods is counted, and the statistical results are respectively divided by the volume of the K first diffusion solder point neighborhoods to obtain the density of the K first diffusion solder point neighborhoods. Specifically, for each first diffusion solder point neighborhood, the number of solder points contained therein is calculated. The volume of each first diffusion solder point neighborhood is re-determined, and the number of solder points in each first diffusion solder point neighborhood is divided by the volume of the neighborhood to obtain the first diffusion solder point neighborhood density.

[0034] Step S524, determining whether the K first diffusion solder point neighborhood densities are greater than or equal to the K initial solder point neighborhood densities, if so, then the K first diffusion solder point neighborhoods are diffused outwardly according to the preset diffusion bandwidth until the preset diffusion times or the preset diffusion stop condition is met, and K target diffusion solder point neighborhoods are obtained, wherein the preset diffusion stop condition is that the difference between the diffusion solder point neighborhood density of the current diffusion solder point neighborhood and the solder point neighborhood density of the last diffusion solder point neighborhood is less than or equal to the preset difference. Specifically, each first diffusion solder point neighborhood density is compared with its corresponding initial solder point neighborhood density. If the first diffusion solder point neighborhood density is greater than or equal to the initial solder point neighborhood density, then continue to diffuse outwardly according to the preset diffusion bandwidth. After each diffusion, the diffusion solder point neighborhood density is recalculated and compared with the solder point neighborhood density of the last diffusion. When the preset diffusion times or the preset diffusion stop condition is met (the difference between the diffusion solder joint neighborhood density of the current diffusion solder joint neighborhood and the solder joint neighborhood density of the previous diffusion solder joint neighborhood is less than or equal to the preset difference), the diffusion is stopped to obtain the target diffusion solder joint neighborhood. Among them, the preset diffusion times is the maximum diffusion times set, which is used to limit the diffusion process. This implementation method ensures that the diffusion process is meaningful by comparing the initial solder joint neighborhood density with the solder joint neighborhood density after diffusion, that is, each diffusion can increase the representativeness or information content of the solder joint neighborhood. By gradually diffusing the initial solder joint neighborhood outward, a suitable solder joint neighborhood range is found, which can contain enough solder joints to reflect the overall performance of the frame solder joints, and will not increase unnecessary detection costs due to containing too many solder joints.

[0035] In a possible implementation, step S530 further includes step S531, removing the K target diffusion solder joint neighborhoods from the frame solder joint abnormality identification topological network, and using the remaining solder joints as multiple discrete solder joints. Specifically, in the frame solder joint abnormality identification topological network, all solder joints in the K target diffusion solder joint neighborhoods are marked. These marked solder joints are removed, and the remaining unmarked solder joints are discrete solder joints, which are solder joints that do not belong to any target diffusion solder joint neighborhood. Step S532, extracting the overlapping solder joints in the K target diffusion solder joint neighborhoods to obtain multiple overlapping solder joints. Specifically, the K target diffusion solder joint neighborhoods are checked to find solder joints that belong to two or more neighborhoods at the same time, extracting these overlapping solder joints, and storing them as a set of overlapping solder joints. Step S533, respectively adding / reserving the multiple discrete solder joints and the multiple overlapping solder joints to the nearest target diffusion solder joint neighborhood to obtain the K adjusted diffusion solder joint neighborhoods. Specifically, for each discrete solder joint, the distance between it and the center of all target diffusion solder joint neighborhoods is calculated, and the discrete solder joint is added to the neighborhood of the target diffusion solder joint closest to it. For each overlapping solder joint, the distance between it and the center of all target diffusion solder joint neighborhoods is also calculated (although it already belongs to multiple neighborhoods, this step is to determine which neighborhood is its "closest"). According to the calculation results, the overlapping solder joint is retained in the original neighborhood (if it is already part of the "closest" neighborhood), or it is moved to another closer neighborhood. After the above processing, the adjusted K diffusion solder joint neighborhoods are obtained. This implementation method ensures that each solder joint is properly classified into a neighborhood by eliminating the neighborhood of the target diffusion solder joint and identifying the discrete solder joints and overlapping solder joints, avoiding omissions or duplications, and improving the accuracy and efficiency of neighborhood division.

[0036] Step S600, traverse the K frame weld abnormality identification sub-topology networks to perform weighted calculation of weld abnormality, and obtain weighted values ​​of weld abnormality of the K sub-topology networks. Specifically, traverse each frame weld abnormality identification sub-topology network, perform weighted calculation on the abnormality of each weld, and the weighted coefficient can be determined based on factors such as the distance from the weld to the center weld, the importance of the weld in the frame, etc. The result of the weighted calculation reflects the comprehensive situation of the weld abnormality in each sub-topology network.

[0037] In a possible implementation, the K frame weld point abnormality identification sub-topological networks are traversed to perform weld point abnormality weighted calculation, and the K sub-topological network weld point abnormality weighted values ​​are obtained. Step S600 further includes step S610, traversing and calculating the distance from each weld point in the K frame weld point abnormality identification sub-topological networks to the K center weld points, and obtaining K weld point distance sets, wherein each weld point corresponds to a weld point distance. Specifically, for each frame weld point abnormality identification sub-topological network (a total of K), traverse each weld point therein. Calculate the distance from the weld point to the center weld point of the sub-topological network to which it belongs, and commonly used distance calculation methods such as Euclidean distance and Manhattan distance can be used here. The calculated distance is stored to form a weld point distance set of the sub-topological network. Repeat this process until K weld point distance sets are obtained, each set corresponding to a sub-topological network.

[0038] Step S620, respectively calculate the ratio of each solder point distance in the K solder point distance sets to the sum of the multiple solder point distances in the corresponding solder point distance set, and obtain K weighted calculation coefficient sets. Specifically, for each solder point distance set (a total of K), traverse each solder point distance therein. Calculate the ratio of the solder point distance to the sum of the sum of all solder point distances in the solder point distance set. This ratio is the weighted calculation coefficient of the solder point, which reflects the relative position weight of the solder point relative to other solder points in the sub-topological network. The calculated weighted calculation coefficients are stored to form a weighted calculation coefficient set of the sub-topological network. Repeat this process until K weighted calculation coefficient sets are obtained, each set corresponding to a sub-topological network.

[0039] Step S630, weighted calculation is performed on the weld abnormality identification corresponding to each weld in the K frame weld abnormality identification sub-topology network according to the K weighted calculation coefficient sets, and the weld abnormality weighted values ​​of the K sub-topology networks are obtained. Specifically, for each frame weld abnormality identification sub-topology network (a total of K), each weld therein is traversed. The weld abnormality identification corresponding to the weld (i.e., the degree of abnormality of the weld) is obtained, and the weighted calculation coefficient corresponding to the weld in the weighted calculation coefficient set of the sub-topology network to which it belongs is searched, and the weld abnormality identification is multiplied by the weighted calculation coefficient to obtain the weighted abnormality of the weld, and the weighted abnormality of all welds is added to obtain the weld abnormality weighted value of the sub-topology network. Repeat this process until the weld abnormality weighted values ​​of the K sub-topology networks are obtained. This implementation method reflects the relative position weight of the weld in the sub-topology network by calculating the distance from each weld to the central weld and determining the weighted calculation coefficient of the weld accordingly. The weighted abnormality of the weld point is obtained by multiplying the abnormality mark of the weld point with the weighted calculation coefficient, which reflects the contribution of the weld point to the overall abnormality. Finally, by calculating the sum of the weighted abnormalities of all weld points in the sub-topological network, the weighted value of the weld abnormality of the sub-topological network is obtained, achieving the technical effect of accurately evaluating the overall abnormality of the frame weld abnormality mark sub-topological network, providing a more accurate reference for subsequent welding performance testing.

[0040] Step S700, according to the size of the weighted value of the abnormal degree of the weld point of the K sub-topological networks and the preset total amount of welding performance detection, K weld point performance detection quantities are determined, and based on the K weld point performance detection quantities, the K frame weld point abnormal degree identification sub-topological networks are sampled for welding performance detection to obtain welding performance detection results. Specifically, the weighted values ​​of the abnormal degree of the weld points of the K sub-topological networks are normalized, and the number of weld points to be detected in each sub-topological network (i.e., weld point performance detection quantity) is obtained by multiplying the normalized result and the preset total amount of welding performance detection. Then, based on these detection quantities, a sampling welding performance detection is performed on each sub-topological network, that is, a certain number of weld points are selected from each sub-topological network for actual performance detection to obtain the final welding performance detection result. The embodiment of the present application adopts technical means such as obtaining basic design information of the target electric three-wheeled sanitation vehicle, searching the historical maintenance database to obtain welding abnormality records and calculate the degree of abnormality of weld points, constructing a frame weld abnormality identification topological network, randomly extracting and authenticating central welds, constructing a sub-topological network, calculating the weighted value of the sub-topological network weld abnormality, determining the detection amount according to the weighted value and performing sampling welding performance detection, etc., to achieve the technical effect of targeted evaluation of frame performance, accurately locating potential welding abnormalities, and improving detection accuracy.

[0041] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for testing the frame performance of an electric three-wheeled sanitation vehicle, characterized in that: The method comprises: Obtaining basic design information of a target electric three-wheeled sanitation vehicle, wherein the basic design information includes the sanitation vehicle model, frame material, and frame welding point topology network; The historical electric three-wheeled sanitation vehicle maintenance database is searched with the sanitation vehicle model, the frame material and the abnormal welding maintenance as indexes to obtain multiple abnormal welding positions and multiple abnormal welding records, and the abnormal welding point degrees are identified for the multiple abnormal welding records to obtain multiple abnormal welding point degrees, wherein each abnormal welding position corresponds to a abnormal welding point degree; Based on the multiple welding abnormality positions and the multiple welding point abnormality degrees, the welding points in the frame welding point topological network are marked with welding point abnormality degrees to obtain a frame welding point abnormality degree marking topological network; Randomly extracting K welding points from the frame welding point abnormality identification topological network as K central welding points, and performing topological network central point authentication on the K central welding points; If the authentication is passed, neighborhood construction is performed in the frame weld abnormality identification topological network according to the K central welds and the corresponding K weld abnormality degrees, to obtain K frame weld abnormality identification sub-topological networks; Traversing the K vehicle frame weld point abnormality identification sub-topological networks to perform weld point abnormality weighted calculation to obtain weld point abnormality weighted values ​​of the K sub-topological networks; K weld performance detection quantities are determined according to the sizes of the weighted values ​​of the weld abnormality degrees of the K sub-topological networks and the preset total welding performance detection amount, and a sampling welding performance detection is performed on the K frame weld abnormality identification sub-topological networks based on the K weld performance detection quantities to obtain a welding performance detection result.

2. The frame performance detection method of the electric three-wheeled sanitation vehicle according to claim 1, characterized in that: Randomly extracting K welding points from the frame welding point abnormality identification topological network as K central welding points, and performing topological network central point authentication on the K central welding points, including: Enumerate the K center welding points in pairs to obtain multiple center welding point combinations; Combined with the frame welding point abnormality identification topological network, distance calculation is performed on the plurality of center welding point combinations respectively to obtain a plurality of center welding point distances; It is determined whether there is a center solder point distance greater than a preset center solder point distance threshold value among the multiple center solder point distances. If so, the authentication result is failed.

3. The frame performance detection method of the electric three-wheeled sanitation vehicle according to claim 2, characterized in that: include: If not, the number of K solder joint anomalies corresponding to the K central solder joints exceeding the preset solder joint anomaly degree is counted to obtain an anomaly amount; When the abnormal amount is greater than the preset abnormal amount threshold, the authentication result is passed.

4. The frame performance detection method of the electric three-wheeled sanitation vehicle according to claim 1, characterized in that: According to the K central welds and the corresponding K weld abnormality degrees, neighborhood construction is performed in the frame weld abnormality degree identification topological network to obtain K frame weld abnormality degree identification sub-topological networks, including: Taking the K central welds as the neighborhood centers, an initial neighborhood is constructed in the frame weld abnormality identification topological network based on a preset diffusion bandwidth to obtain K initial weld neighborhoods, wherein each initial weld neighborhood includes a plurality of welds whose distance to the central weld is less than or equal to a preset diffusion bandwidth; Performing neighborhood diffusion identification on the K initial solder joint neighborhoods according to the preset diffusion bandwidth to determine K target diffusion solder joint neighborhoods; Performing neighborhood adjustment on the K target diffusion solder joint neighborhoods based on the frame solder joint abnormality identification topological network to obtain K adjusted diffusion solder joint neighborhoods; The K frame weld point abnormality identification sub-topological network is constructed according to the positions of each weld point in the neighborhood of the K adjusted diffusion weld points.

5. The frame performance detection method of the electric three-wheeled sanitation vehicle according to claim 4, characterized in that: include: Counting the number of welds in the K initial weld neighborhoods, and respectively dividing the statistical results by the volumes of the K initial weld neighborhoods to obtain the density of the K initial weld neighborhoods; Diffusion is performed on the K initial solder joint neighborhoods outward based on the preset diffusion bandwidth to obtain K first diffusion solder joint neighborhoods; Counting the number of solder joints in the neighborhoods of the K first diffusion solder joints, and respectively dividing the statistical results by the volumes of the neighborhoods of the K first diffusion solder joints to obtain the neighborhood densities of the K first diffusion solder joints; Determine whether the density of the K first diffusion solder point neighborhoods is greater than or equal to the density of the K initial solder point neighborhoods. If so, diffuse the K first diffusion solder point neighborhoods outward according to the preset diffusion bandwidth until a preset diffusion number or a preset diffusion stop condition is met, and obtain K target diffusion solder point neighborhoods, wherein the preset diffusion stop condition is that the difference between the diffusion solder point neighborhood density of the current diffusion solder point neighborhood and the solder point neighborhood density of the previous diffusion solder point neighborhood is less than or equal to the preset difference.

6. The frame performance detection method of the electric three-wheeled sanitation vehicle according to claim 5, characterized in that: include: The K target diffusion welding point neighborhoods are removed from the frame welding point abnormality identification topological network, and the remaining welding points are used as a plurality of discrete welding points; Extracting overlapping solder joints in the neighborhood of the K target diffusion solder joints to obtain a plurality of overlapping solder joints; The plurality of discrete solder joints and the plurality of overlapping solder joints are respectively added / reserved to the nearest target diffusion solder joint neighborhood to obtain the K adjusted diffusion solder joint neighborhoods.

7. The frame performance detection method of the electric three-wheeled sanitation vehicle according to claim 1, characterized in that: Traversing the K frame weld point abnormality identification sub-topology networks to perform weld point abnormality weighted calculation to obtain K sub-topology network weld point abnormality weighted values, including: Traversing and calculating the distance between each welding point and the K central welding points in the K frame welding point abnormality identification sub-topological network, and obtaining K welding point distance sets, wherein each welding point corresponds to a welding point distance; Calculate the ratio of each welding point distance in the K welding point distance sets to the sum of multiple welding point distances in the corresponding welding point distance set to obtain K weighted calculation coefficient sets; A weighted calculation is performed on the weld point abnormality identification corresponding to each weld point in the K frame weld point abnormality identification sub-topology networks according to the K weighted calculation coefficient sets to obtain weld point abnormality weighted values ​​of the K sub-topology networks.

8. The frame performance detection method of the electric three-wheeled sanitation vehicle according to claim 1, characterized in that: include: Acquire multiple sample welding abnormality records and multiple sample welding point abnormality degrees as training data; Using the training data to supervise the framework built based on the BP neural network, learning the mapping relationship between welding abnormality records and welding point abnormality, until the training converges, and obtaining a trained abnormality recognition network layer; The abnormality degree identification network layer is used to identify the abnormality degrees of weld points on the multiple welding abnormality records to obtain the abnormality degrees of the multiple weld points.

Images