WO2019207952A1

WO2019207952A1 - Vibration noise response plan recommendation system

Info

Publication number: WO2019207952A1
Application number: PCT/JP2019/007676
Authority: WO
Inventors: 武藤　大輔; 直也三津橋; 源太山内; 吉澤　尚志; 洋祐田部; 真理黒澤; 高野　靖
Original assignee: 株式会社日立製作所
Priority date: 2018-04-24
Filing date: 2019-02-27
Publication date: 2019-10-31
Also published as: JP2019191851A

Abstract

Provided is a vibration noise response plan recommendation system with which it is possible to present cost effectiveness for each option of measures for suppressing vibration or noise specifically when designing a product to newly develop or examining plans for improving an existing product. A vibration noise response plan recommendation system 1 is provided with: a cause-of-vibration-noise problem estimation device 30 for estimating the vibration noise propagation state of actual operation state of an apparatus and the rate of contribution thereof; a vibration-noise problem cost-effectiveness determination device 40 for generating a vibration-noise problem suppression probability table 32, as well as generating a cost table 34 on the basis of a prior-subsequent cost list 33 comprising the prior cost required for the implementation of a suppression measure for each of inputted vibration noise problems and the subsequent cost required after the occurrence of a problem, and generating a pay-off table 35 pertaining to a cost expectation value on the basis of the vibration-noise problem suppression probability table 32 and the cost table 34; and a design data management and storage unit 59 for storing the prior-subsequent cost list 33 and a game tree 50, which is a tree structure of inputted action selection and status determination in each process of product design and manufacture. The vibration noise response plan recommendation system 1 generates the pay-off table 35 pertaining to the cost expectation value of each option of action selection in the game tree 50.

Description

Vibration noise countermeasure plan recommendation system

The present invention relates to a system for determining cost effectiveness related to implementation of a proactive measure for preventing the occurrence of a vibration noise problem, and in particular, a vibration noise countermeasure plan suitable for decision support for implementing a proactive measure based on the cost effectiveness. Regarding the recommended system.

∙ Vibration and noise generated from the product are widely recognized as factors that deteriorate the comfort when using the product. For this reason, it is important to understand and control vibration noise.To this end, it is necessary to grasp the source and transmission path of vibration and noise in advance at the design stage, and based on the results, change the current structure plan and Prior study is required. By the way, in general, anti-vibration / sound-proof measures implemented to suppress vibration noise not only increase cost and weight but also have a trade-off relationship with other performance. For this reason, it is possible to avoid implementing measures aimed solely at suppressing vibration and noise, and to optimize the strategy of taking post-measures after a problem occurs without applying anti-vibration / sound-proof measures in advance. If so, you can choose. On the other hand, it must be avoided that the cost of subsequent measures taken after a problem occurs is higher than the expected cost when measures are taken in advance.

In this way, ultimately, the required specifications for vibration and noise can be achieved at the barely necessary level, and the total cost including the cost of measures, weight, and other effects on performance can be minimized. Although it is necessary to apply measures that can be realized, it is actually difficult to judge. This is because there are uncertainties in various factors related to performance and cost in the initial design stage, and it is difficult to make deterministic judgments. Also, there are large variations in construction conditions and material characteristics during actual construction. It is.

Regarding the technology for decision support regarding such precautions against vibration noise, for example, as described in Patent Document 1, data necessary for calculation so that even a less experienced user can handle noise prediction around a site / building. Has been proposed to display the noise spectrum and the contribution ratio for each source, by making it easier to input by reading a map of the site and building. Further, as described in Patent Document 2, a technique for estimating the cause and factors of noise based on uncertain information and past countermeasure examples has been proposed. Further, as disclosed in Patent Document 3, regarding noise source estimation, a technology for determining whether a sound source whose noise type is not specified is a sound source to be analyzed using a probability density distribution. Proposed.

JP 2006-260370 A JP-A-10-187784 Republished patent WO2009 / 139052

In order to suppress the vibration and noise generated by the product, it is necessary to first understand the sound source, excitation source and transmission path, and in general, the sound source and transfer characteristics are identified using measured data. Further, the dominant transmission path and the contribution rate indicating the magnitude of the contribution from each path are determined.

As described above, in general, the anti-vibration / sound-proofing measures implemented to suppress vibration noise not only increase the cost and weight but also have a trade-off relationship with other performance. For this reason, it is necessary to apply measures that can achieve the specifications related to vibration and noise at a marginal and necessary level, and that can be realized while minimizing the total cost including the cost of the measures, the weight, and the effect on other performance. However, that judgment is actually difficult.

The difficulty of this judgment will be explained using drawings. FIG. 10 is a diagram for explaining a problem in the determination of whether or not the conventional measures against vibration and noise can be implemented. As shown in the cost table 34 in the upper diagram of FIG. 10, it is assumed that there is a certain noise and noise reduction measure. If this is applied in advance at the initial design stage, the construction cost of the measure, redesign or adjustment of other performance, etc. It is assumed that the cost including the cost (preliminary cost) is “2”. In addition, if it is determined that vibration and noise do not reach the target performance after product production / shipment, it is assumed that a cost of “5” (ex-post cost) will be incurred due to countermeasures, design changes or payment of a penalty to the customer. . Then, the cost table 34 is as shown in the lower diagram of FIG. The most promising is when there is no problem after the fact even if no prior action is taken, and the total cost at this time is “0” (best condition). On the other hand, what should be avoided the most is a case where a problem occurs after the fact despite the fact that it has been handled in advance, and the total cost at this time is “7” (worst condition). Now, suppose that there is no causal relationship between the implementation of proactive measures and the occurrence of post-problems. In this case, if it is known that a problem does not occur after the fact, it is better not to implement a proactive measure (total cost: 0) than to implement a proactive measure (total cost: 2). Similarly, if it is known that a problem will occur after the fact, it is better not to implement a proactive measure (total cost: 5) than to implement a proactive measure (total cost: 7). In any case, it is better not to implement a proactive measure. In game theory, this is called a “domination strategy”. If you do not recognize a causal relationship between the implementation of proactive measures and the occurrence of subsequent problems, it is always better not to implement proactive measures at a lower cost and better strategy. It is judged that.

On the other hand, as long as it is a proactive measure aimed at preventing problems, there is a causal relationship between the implementation of proactive measures and the occurrence of subsequent problems. If the proactive measures are implemented, the posterior problems are surely suppressed. Conversely, if the proactive measures are not implemented, the posterior problems will surely occur. FIG. 11 shows the vibration noise problem suppression accuracy table 32 and the cost table 34 in this case, and “2” is the first to show the causal relationship between the implementation of such a proactive measure and the occurrence of a post-problem. Determining the occurrence of subsequent problems by investing costs and implementing proactive measures is determined to be better than paying the cost of “5” for countermeasures for subsequent problems without implementing proactive measures. I can do it.

However, in the measurement of vibration noise phenomenon, generally, there are many cases in which data with variations is obtained due to noise mixing and subtle fluctuations in measurement conditions. Therefore, the above-described sound source and transfer characteristic identification results also include a lot of uncertainties due to data variations. In other words, the assumed transmission path and vibration / noise contribution rate determined based on the transmission model based on such actual measurement results, and the prospects for structural changes calculated based on it are naturally deterministic. It is thought that it is not possible to make an argument That is, it is difficult in the actual problem to explain with a certain causal relationship as shown in FIG. For this reason, it is desirable that probabilistic treatment is performed in consideration of variations and uncertainties, and the decision support for implementing structural changes derived from the results is also based on the probability distribution of the vibration noise estimate and the degree of its influence. It is desirable to make a determination based on it. Here, for convenience of explanation, only two values, “implement” and “not implement”, are handled as measures, but originally “do not implement”, “implement measure A”, and “implement measure B” In general, there are three or more measures / options. In addition, there are only two values for the occurrence of a subsequent problem: “occurs after the problem occurs” and “does not occur after the problem”. However, the cost of the subsequent countermeasures generally changes depending on the amount of the target not achieved. Therefore, it will generally be at multiple levels.

By the way, until the product is completed, several inspections, evaluations, and tests are performed in order to confirm the reliability of the product in each manufacturing process. Therefore, the reliability of this estimation gradually increases by reflecting the results of these inspection tests in the estimation regarding whether or not the vibration noise target value of the product can be achieved. For example, FIG. 12 is a diagram showing a case of vibration noise evaluation of a railway vehicle. In this case, first, an element test 52 for confirming whether the countermeasure partial structure has a predetermined performance is performed at the initial stage of design. It is. In the latter half of the design, a structure test of a prototype vehicle to which the vibration and noise countermeasure partial structure is applied is performed, and here, it is confirmed whether or not the assumed performance can be obtained as a whole structure. After the design is completed, production will start. Various tests are conducted here, and after the final judgment of the specification achievement prospect is finally made in the in-place stationary test of the finished vehicle, the judgment of achievement of the specification is made by the running test. Made. In this way, by reflecting the evaluation / test data generated as the design and production process 51 progresses, the causal relationship between the implementation of the proactive measures shown in FIG. It is common to think that sex gradually increases. In other words, the probability of occurrence of the posterior problem for the proactive measures as shown in FIG. 11 is gradually 0% or 100% by making good use of the inspection / evaluation test results conducted through the product manufacturing process as shown in FIG. In this process, if it is judged that the target has not been achieved, it is necessary to review the design judgment at any time in view of cost effectiveness. Although not shown, it is conceivable that the data at the time of operation and maintenance is monitored after shipment and the obtained data is reflected in the design of the next model again.

As such a vibration noise prediction calculation technique, for example, there is a technique described in Patent Document 1, and there is a description regarding a technique for predicting vibration noise under a given condition based on stored and managed design data. There is, however, no treatment for probabilistic reasoning about uncertain factors. Further, as a technique having probabilistic reasoning, it is specified whether or not this is an analysis object from the technique described in Patent Document 2 inferred from past cases, or measurement data of a sound source for which the type of noise is not specified. For this reason, there is Patent Document 3 which presents a method for inferring, but neither of them has a technology for presenting cost-effectiveness related to design options.

As described above, it expresses the uncertainty of various conditions such as the technology to predict vibration noise under the given conditions based on the stored and managed design data, the characteristics of the sound source / excitation source and the transmission characteristics of the transmission path However, there is a technology that presents the best options. However, in Patent Document 1 to Patent Document 3 described above, although there is a trade-off relationship between the vibration noise countermeasure and the cost required for it, the cost effectiveness is not presented for each option of various vibration noise countermeasures.

Therefore, the present invention provides a vibration and noise countermeasure plan that can present a cost-effectiveness for each option of measures for suppressing vibration or noise, particularly when designing a newly developed product or considering an improvement plan for an existing product. Provide a recommended system.

In order to solve the above-mentioned problems, a vibration noise countermeasure plan recommendation system according to the present invention includes (1) a vibration noise problem factor estimating device for estimating a vibration noise transmission state and a contribution rate of an actual operating state of the device, and (2) Based on the vibration noise transmission path probability reasoning model generated by the vibration noise problem factor estimation device, a vibration noise problem suppression accuracy table that correlates the presence / absence of implementation of a proactive measure and a post-problem occurrence suppression rate is generated and each input A cost table is generated based on a prior cost list consisting of a prior cost for implementing a measure for suppressing vibration noise problems and a subsequent cost after the occurrence of the problem, and a gain related to expected cost based on the vibration noise problem suppression accuracy table and the cost table Vibration and noise problem cost-effectiveness judgment device that generates a table, and (3) action selection in each process of design and production of input products A game tree that is a tree structure for situation determination; and a design data management / storage unit that stores the pre- and post-cost list, and the vibration noise problem cost-effectiveness determination device is configured to select each of action selections in the game tree. It is characterized by generating and presenting to the user a gain table relating to the expected cost of the option.

According to the present invention, a vibration and noise countermeasure plan capable of presenting a cost-effectiveness for each option of a measure for suppressing vibration or noise, particularly when designing a newly developed product or considering an improvement plan for an existing product. It is possible to provide a recommended system.
Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

It is a functional block diagram of a vibration noise problem cost-effectiveness determination apparatus according to an embodiment of the present invention. It is a figure explaining action selection and situation judgment performed based on data obtained in a product design / production process. It is a figure explaining the process outline | summary of the vibration noise problem cost-effectiveness determination apparatus shown in FIG. FIG. 4 is a diagram illustrating cost-effectiveness determination regarding the implementation of vibration and noise countermeasures illustrated in FIG. 3 together with action selection and situation determination during the design / production process illustrated in FIG. 2. It is a figure explaining the structural example of the vibration noise problem factor estimation apparatus which comprises the vibration noise countermeasure plan recommendation system which concerns on one Example of this invention. It is a figure which shows the example of a screen display of the contribution rate display part which comprises the vibration noise problem factor estimation apparatus shown in FIG. It is a whole schematic block diagram of the vibration noise countermeasure plan recommendation system which concerns on one Example of this invention, Comprising: It is a figure explaining data and a process. FIG. 5 is a diagram for explaining that the selection criteria for action selection and situation determination during the design / production process shown in FIG. 4 change depending on the discovery of a new solution measure. It is a figure explaining the situation determination as a result of the action selection in the design and production process shown in FIG. 8, and the recommended strategy of the next action selection in the present condition on that. It is a figure explaining the problem in judgment of the conventional vibration noise countermeasure implementation feasibility. It is a figure explaining the other problem in judgment of the conventional vibration noise countermeasure implementation feasibility. It is a figure explaining the case of a railroad vehicle about the data extraction work, such as a design / manufacturing process of a product, the test | inspection performed in the process, evaluation, and a test, and the data obtained by these data extraction work.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a functional block diagram of a vibration / noise problem cost-effectiveness determination apparatus according to an embodiment of the present invention, and FIG. 2 is an action selection and situation determination performed based on data obtained in a product design / production process. FIG. 3 is a diagram for explaining the processing outline of the vibration noise problem cost-effectiveness determination apparatus shown in FIG.
The conventional method shown in FIG. 11 is based on the premise that there is a certain causal relationship between the implementation of the proactive measures and the occurrence of the posterior problem, and takes into account variations and uncertainties that occur in the vibration noise problem. The problem was that there was no probabilistic treatment. Further, a systematic mechanism for updating the next measure application strategy using the data 53 obtained sequentially as the design and production process 51 progresses as shown in FIG. 12 has not been incorporated. FIG. 2 is a diagram for explaining the application of the data 53 obtained sequentially by performing the evaluation test 52 to the design judgment as these steps progress. FIG. 2 shows a hierarchical chain structure of a tree that performs conditional branching related to situation determination 55 from evaluation test data 53 of each process obtained as a result of conditional branching related to design selection action selection 54 and each process. What is recorded as a history together with its rationale is stored and / or updated in the design data management / storage unit 59 described later in detail as knowledge 56 in each process. As a result, it is possible to visually explain where the current situation is in the current past and future situations.
In the development type game theory, such a branch diagram of action selection and situation determination in the design and production process is referred to as a game tree 50. What is important at this time is the judgment reasoning for the action selection 54.

In the vibration noise problem cost-effectiveness determination apparatus 40 according to the present embodiment shown in FIG. 1, the uncertainties of the causal relationship that could not be handled in FIG. 11 described above are incorporated as probabilities. As shown in FIG. 1, the vibration noise problem cost-effectiveness determination device 40 includes a cost table generation unit 41, a vibration noise problem suppression accuracy table generation unit 42, a gain table generation unit 43, an input I / F 44, and an output I / F 45. Is provided. The cost table generation unit 41, the vibration noise problem suppression accuracy table generation unit 42, and the gain table generation unit 43 include, for example, a processor (not shown) such as a CPU (Central Processing Unit), a ROM that stores various programs, and data of calculation processes. Is realized by a storage device such as a RAM and an external storage device, and a processor such as a CPU reads out and executes various programs stored in the ROM, and the operation result as an execution result is stored in the RAM or the external storage. Store in the device. At least one of the cost table generation unit 41, the vibration noise problem suppression accuracy table generation unit 42, and the gain table generation unit 43 constituting the vibration noise problem cost-effectiveness determination device 40 is, for example, a program (for example, in a program memory). The program may be realized by a stored program. In other words, all of the cost table generation unit 41, the vibration noise problem suppression accuracy table generation unit 42, and the gain table generation unit 43 may be incorporated into one program, or a desired combination may be incorporated into one program. good.

As shown in FIG. 3, the post-problem occurrence suppression rate when the proactive measure for the vibration noise problem is implemented is 80%, and the post-problem occurrence suppression rate when the proactive measure for the vibration noise problem is not implemented. Is assumed to be 30%. Then, the gain table 35 at the time of making a strategy becomes an expected cost value, which is obtained by multiplying the cost table 34 by the vibration noise problem suppression accuracy table 32. From this gain table 35, it can be explained that the expected cost value when implementing the proactive measure is “3.0”, which is a better strategy than the expected cost value “3.5” when not implementing the proactive measure. I can do it. As will be described in detail later, the cost table generation unit 41 constituting the vibration noise problem cost-effectiveness determination device 40 has a function of generating the cost table 34, and the vibration noise problem suppression accuracy table generation unit 42 suppresses the vibration noise problem. A function to generate the accuracy table 32 is provided. Further, the gain table generation unit 43 multiplies the vibration noise problem suppression accuracy table 32 generated by the vibration noise problem suppression accuracy table generation unit 42 by the cost table 34 generated by the cost table generation unit 41, thereby obtaining a gain table. 35 is generated.
By using this concept, the overall cost, including the cost, weight, and other performance impacts, is minimized even for cost-effectiveness of highly variable data and highly uncertain measures. Thus, it is an effective method for determining an optimum measure when it is desired to achieve a required specification at a barely sufficient level.

FIG. 4 shows the gain table 35 generated by the gain table generation unit 43 constituting the vibration noise problem cost-effectiveness determination apparatus 40 described in FIG. 3 together with the game tree 50 of FIG. 2 described above.
FIG. 4 shows that the respective situation judgments 55 are executed from the test evaluation data 53 obtained in the step 51 at that time with respect to the results obtained as the results of the respective options in the action selection 54. The occurrence probability and cost of each event at each end of the game tree 50, and the expected cost value obtained by multiplying them are shown as a gain table 35. The above-described processing is in principle the same as the calculation of the gain table 35 from the vibration noise problem suppression accuracy table 32 and the cost table 34 shown in FIG. 3, and the process of creating the gain table 35 shown in FIG. Thus, the vibration noise problem cost-effectiveness determination apparatus 40 shown in FIG. 1 operates. In FIG. 4, the name of the measure or the situation is shown above each arrow derived from the action selection 54 and the situation judgment 55, and the [] below each arrow is generated by the action selection 54 or the response to the situation. The cost to perform (in this case, the recovery cost when the target is not achieved) is represented by a numerical value. Here, as an example, among the three options “Measure A0 (No Policy Implementation: Cost 0)”, “Measure A1”, and “Measure A2” that can be taken in the initial action selection 54, Judging from the prospect of the effect, although the current cost of “Measure A1” is “2”, it is the most appropriate with the smallest expected cost in the future (in the case of FIG. 4, the expected cost of each judgment is “3.0”) is shown along with the game tree 50 as a reason for determining that “3.0”). In addition, recording and leaving such a game tree 50 and the gain table 35 as a judgment rationale at a certain design stage is important knowledge when designing the next model later.

In FIG. 4, the name of each measure or situation derived from the action selection 54 and the situation judgment 55 is input and set by the user (designer) at the design stage. Further, the vibration noise problem cost-effectiveness determination device 40 calculates an expected cost value of each option for each measure based on a prior and subsequent cost list 33 input by a user (designer) described later in detail. In other words, in the game tree 50, the policy and the situation itself branching in each design / production process (each level) 51 are set by the user (designer), but the policy branches in each design / production process (each level) 51. In addition, in order to comprehensively evaluate the combination of the total measures including the situation and the prospects (achieved prospects, unachieved prospects), the vibration and noise problem cost-effectiveness judging device 40 is expected to provide the cost expected value of each measure option, that is, each judgment Calculate the expected cost of. The action selection 54 means extraction of measure options, for example.

FIG. 5 is a diagram for explaining a configuration example of a vibration noise problem factor estimation device constituting a vibration noise countermeasure plan recommendation system according to an embodiment of the present invention. What is important in obtaining the game tree 50 and the gain table 35 based thereon shown in FIG. 4 is the vibration noise problem suppression accuracy table generation unit 42 (FIG. 1) constituting the vibration noise problem cost-effectiveness determination device 40. FIG. 3 is a vibration noise problem suppression accuracy table 32 (FIG. 3) generated by The vibration noise problem factor estimation device 30 shown in FIG. 5 is necessary for constructing the vibration noise problem suppression accuracy table 32. Hereinafter, the configuration of the vibration noise problem factor estimation device 30 will be described.
As shown in FIG. 5, the vibration noise problem factor estimation device 30 includes an on-site data extraction unit 11 that extracts various types of vibration and noise data in the actual operation state of the existing measurement target product 10, and an on-site data extraction unit 11. An on-site data storage unit 12 is provided for storing the extracted actual operation data on the site. Further, the vibration noise problem factor estimating device 30 generates a transmission path model 16a in which the vibration noise transmission path of the measurement target product 10 estimated by the user (designer) 20 and the transmission path is expressed by a graphical model. A laboratory data storage unit 17 to be stored in the model data storage unit 13, a sound source / transfer characteristic estimation engine 14, a sound source / transfer characteristic probability distribution data storage unit 15, a contribution rate estimation engine 18, and a contribution rate display unit 19 are provided. The sound source / transfer characteristic estimation engine 14, the laboratory data storage unit 17, and the contribution rate estimation engine 18 are, for example, a processor (not shown) such as a CPU (Central Processing Unit), a ROM for storing various programs, and data of calculation processes temporarily. Is realized by a storage device such as a RAM and an external storage device, and a processor such as a CPU reads out and executes various programs stored in the ROM, and the operation result as an execution result is stored in the RAM or the external storage device. Store.

The sound source / transfer characteristic estimation engine 14 includes a plurality of actual operation site data stored in the site data storage unit 12 and a transfer path model stored in the transfer route model data storage unit 13 and selected by the user (designer) 20. 16a, the probability distribution of the sound source / transfer characteristic corresponding to each transfer path of the transfer path model 16a is estimated. The sound source / transfer characteristic estimation engine 14 stores the estimated probability distribution of the sound source / transfer characteristic in the sound source / transfer characteristic probability distribution data storage unit 15.
The contribution rate estimation engine 18 includes a transmission path model 16a stored in the transmission path model data storage unit 13, a sound source / transfer characteristic probability distribution data stored in the sound source / transfer characteristic probability distribution data storage unit 15, and a field data storage unit. The vibration noise contribution rate is calculated on the basis of a plurality of actual operation site data stored in 12. The contribution rate estimation engine 18 outputs the calculated vibration noise contribution rate to the contribution rate display unit 19. The user (designer) 20 operates the contribution rate display unit 19 to display the contribution rate with respect to a predetermined partial structural element of the measurement target product 10 on a display screen (not shown) of the contribution rate display unit 19. In addition, it is possible to obtain a probabilistic inference result regarding the degree of influence on the entire measurement target product and the cost effectiveness when planning a structural change.
Note that the sound source / transfer characteristic probability distribution data 16b acquired in the pre-shipment test or the periodic inspection test of the measurement target product 10 is generated by the user (designer) 20 by the laboratory data storage unit 17 by the sound source / transfer characteristic probability distribution. It is also possible to store and update the data storage unit 15.

As described above, the laboratory data storage unit 17 allows the user (designer) 20 to store the transmission path model 16a and the sound source / transfer characteristics that can be stored in the transmission path model data storage unit 13 and the sound source / transfer characteristic probability distribution data storage unit 15, respectively. The probability distribution data 16b can be examined and calculated based on laboratory measurement results and analysis results, thereby updating and improving the vibration noise model of the product to be examined. For this reason, the transfer path model 16a and the sound source / transfer characteristic probability distribution data 16b are collectively referred to as laboratory data 16 here.

As an algorithm for the probabilistic estimation by the above-described sound source / transfer characteristic estimation engine 14, for example, a technique called a Bayesian network is used. A Bayesian network assumes a causal relationship between an event H and data D obtained thereunder, and a conditional probability density distribution in which the data D occurs under the occurrence of the event H is a likelihood P (D | H). And the probability of occurrence of the event H from which the data D ₁ can be obtained when the data D ₁ is obtained at the actual site is the likelihood P (D of the data D ₁ for the event H shown above. ₁ | H) is an estimation method.

For example, when the vibration noise S _i at the evaluation point and the signal such as the excitation force F or the sound source N can be directly measured in an actual operation test or the like, first, in the learning phase, the transfer characteristic H ₁ = S _i / F between them and H ₂ = S _i / N is calculated backward, and as a result, the conditional probability density distribution of the vibration noise S _{i at} the evaluation point for H ₁ and H _{2 is} converted into a database as likelihood P (S _i | H ₁ ∩H ₂ ). And stored in the sound source / transfer characteristic probability distribution data storage unit 15. Then, the estimated phase, after obtaining the vibration noise _{S n} of the evaluation point at the time of actual operation, the transmission characteristic data can be observed _H 1, likelihood relates _{_{_{H 2 P (S n | H}}} 1 ∩H 2) Are sequentially multiplied by the number of observation signals, and the transfer characteristics H ₁ and H ₂ during actual operation can be identified probabilistically.

It should be noted that the sound source and the transfer characteristic are interchangeable in that they are multiplied by the vibration noise at the evaluation point. Therefore, if the above method is taken in reverse, it can be used for identification of excitation sources and sound sources that are difficult to measure directly during actual operation. For example, an arbitrary excitation force F or likelihood P (S _i | F∩N) between the sound source N and the vibration noise S _{i at} the evaluation point is acquired in a pre-shipment test or the like, and this is converted into a database. It is stored in the transfer characteristic probability distribution data storage unit 15. Next, the probability density P (S _n | F∩N) of the excitation force F or the sound source N that can be observed from this vibration noise S _n at the evaluation point acquired during actual operation is the number of observation signals. By sequentially multiplying only, the excitation force and the size of the sound source during actual operation can be identified stochastically.

In this way, by estimating the excitation force / sound source characteristics and transfer characteristics in the current model, the contribution of the evaluation point vibration noise in the current model is probabilistically calculated, and based on this, the specifications and design of the next model are calculated. Guideline can be considered.
After that, when the next model is completed, in the pre-shipment test, the likelihood data of the next model's excitation force and the transmission characteristics of the evaluation point vibration noise for the sound source are acquired, and the excitation force / sound source of the current model described above is obtained. By multiplying the probabilistic identification results, the evaluation point vibration noise during operation of the next model can be predicted probabilistically in advance before operation. The above-mentioned evaluation points are basically set by the user (designer) 20, but cannot be set as evaluation points for portions where measured data cannot be measured by sensors or the like.

FIG. 6 is a diagram showing a screen display example of the contribution rate display unit 19 constituting the vibration noise problem factor estimating device 30 shown in FIG. As shown in FIG. 6, the display screen of the contribution rate display unit 19 displays a transfer path graphical model display area 19 a that displays the transfer path model indicated by the graphical model, and a two-dimensional graph that represents the frequency characteristics of the contribution rate. A contribution rate two-dimensional graph display area 19b and an influence / probability determination element display area 19c for displaying design determination materials are configured. When a desired element is selected by a mouse pointer or the like on the graphical model representing the transmission path displayed in the transmission path graphical model display area 19a, variations in sound source / transfer characteristics related to the element are included in the overall estimation result. The degree of influence is displayed on the two-dimensional graph in the contribution rate two-dimensional graph display area 19b. Specifically, as shown in FIG. 6, from the variation of the sound source / transfer characteristics related to the element selected in the transfer path graphical model display area 19a, the upper limit and lower limit of the contribution ratio of vibration noise passing through the element, or The confidence interval is displayed in the contribution rate two-dimensional graph display area 19b. Based on this figure, it is possible to determine the influence of the variation in the vibration noise characteristics of the selected elements on the vibration noise characteristics of the entire system, specifically the degree and probability of upward and downward vibrations. For example, in FIG. 4, the frequency characteristic of the evaluation point noise is indicated by a thick black line in the contribution rate two-dimensional graph display area 19b, and the median of the contribution rate of the selected element is indicated by a gray line, and the variation is represented by a confidence interval. (Vertical bar), and the fluctuation width of the evaluation point noise due to the uncertainty of the contribution rate is indicated by a gray broken line. As a result, the possibility / risk judgment element display area 19c shows the possibility that the evaluation point noise will increase by 2 dB or more with a probability of 25% and decrease by 2 dB or more with a probability of 5%.
Further, if such an influence degree calculation result is further processed, it is possible to provide a function for displaying structural elements that need to be examined in order to suppress vibration noise at the evaluation point in order of priority of countermeasures.

FIG. 7 is an overall schematic configuration diagram of the vibration and noise countermeasure plan recommendation system 1 according to the present embodiment, and is a diagram for explaining data and processing.
As shown in FIG. 7, the vibration noise countermeasure plan recommendation system 1 includes the vibration noise problem factor estimation device 30 shown in FIG. 5, the vibration noise problem cost effectiveness determination device 40 shown in FIG. 1, and design data management. A storage unit 59 is included. When the vibration noise problem factor estimation device 30 automatically generates the vibration noise transmission path probability inference model 31, the vibration noise problem suppression accuracy table generation unit 42 constituting the vibration noise problem cost effectiveness determination device 40 generates the generated vibration noise transmission. Based on the path probability inference model 31, a vibration noise problem suppression accuracy table 32 is generated. The generation of the vibration noise problem suppression accuracy table 32 is not generated by the vibration noise problem suppression accuracy table generation unit 42 based on the vibration noise transmission path probability inference model 31, but is generated. The vibration noise problem suppression accuracy table 32 may be automatically generated based on the model 31.
The user (designer) 20 has a game tree 50 that is a tree structure of action selection and situation determination in the product design and production process, and prior costs for implementing measures for suppressing each vibration and noise problem in each action selection, and occurrence of problems. When the subsequent cost is estimated and input as the prior and subsequent cost list 33, these are temporarily stored in the design data management / storage unit 59. The cost table generation unit 41 constituting the vibration noise problem cost-effectiveness determination apparatus 40 inputs the game tree 50 and the pre-post and post-cost list 33 input to the design data management / storage unit 59 via the input I / F 44. Then, the cost table generation unit 41 generates a cost table 34 for each option in each action selection 54 (FIG. 4) based on the game tree 50 and the pre- and post-cost list 33. The gain table generation unit 43 constituting the vibration noise problem cost effectiveness determination device 40 is generated by the vibration noise problem suppression accuracy table 32 generated by the vibration noise problem suppression accuracy table generation unit 42 and the cost table generation unit 41. Based on the cost table 34, a gain table 35 for each option in an action selection 54 in the process is generated. The vibration noise problem cost-effectiveness determination device 40 outputs the output I / O including the optimum applied measure in the action selection 54 and the expected vibration noise level and contribution rate when the measure is implemented, including the variation and the fluctuation range thereof. It outputs to the contribution rate display part 19 which comprises the vibration noise problem factor estimation apparatus 30 via F45. Thereby, on the display screen of the contribution rate display unit 19, for example, the display information shown in FIG. 6 described above, in other words, at the time of designing a product to be newly developed or at the time of examining an improvement plan of an existing product, Evaluation information for measures for suppressing vibration or noise can be easily visually confirmed. As for the display on the contribution rate display unit 19, a contribution rate two-dimensional graph for option judgment in a certain design / production process 51 as shown in FIG. 6 may be displayed, and in the overall process 51 as shown in FIG. The game tree 50 and the gain table 35 may be displayed side by side. In addition, when the game tree 50 and the gain table 35 are displayed side by side as shown in FIG. 4, the recommended strategy with the smallest cost expectation value at that time is highlighted in a chained manner (from the most upstream to the most downstream (end) of the design and production process 51. ) Is preferably highlighted (highlighted or otherwise distinguishable)). By doing in this way, it becomes possible to look down on the present situation, future prospects, and optimum strategy in the whole process.

FIG. 8 is a diagram for explaining that the selection criteria for action selection and situation judgment during the design / production process shown in FIG. 4 change depending on the discovery of a new solution measure. That is, FIG. 8 shows, as an example, how the initial strategy is changed by finding a new discovery solution measure 57 as a recovery measure for achieving the target in the previous process. Specifically, when the measure A0 of the cost “0” is performed in the process A under the situation shown in FIG. 4, as a response method when the target is not reached in the evaluation test in the process A, This shows the situation after the prospect of achieving the target with a relatively high probability by implementing the measure B1 that pays the additional cost “2” in the measure in the process B. The expected cost value when applying the strategy (measure A0) is “2.4”, which is smaller than the expected cost value “3.0” when the measure A1 is implemented in the initial process A. Is shown. That is, by finding the above-mentioned new discovery solution measure 57, it is determined that there is no problem in process A because even if measure A0 that is less costly than measure A1, which is the initial strategy, can be recovered later. It will be updated. As described above, in the present embodiment, the optimum judgment using the latest knowledge at the present time can be shown by reflecting the estimation result of the problem suppression accuracy of the examination measure at that time as needed and reviewing the gain table 35. .

FIG. 9 is a diagram for explaining a situation determination as a result of action selection during the design / production process shown in FIG. 8 and a recommended strategy for the next action selection in the current situation. That is, FIG. 9 shows the situation where the target is not reached after actually implementing the measure A0 in the situation shown in FIG. 8 above, and further investigates the vibration noise problem suppression accuracy of the measure B1 at that time. As a result, the situation worsened from the initial assumption shown in FIG. As shown in FIG. 9, the expected cost value at the time of implementation of the measure B1 at this time increases to “5.5” due to the above-described deterioration of the vibration noise problem suppression accuracy, and the implementation cost “0” that the target cannot be achieved is inevitable. After implementing the measure B0, it is determined that it is better to pay the post-cost “5.0” for not achieving the target. In this case, later, because the prospect of the vibration noise problem suppression accuracy of measure B1 that was initially assumed was unsatisfactory, an erroneous determination was made that “measure A0 was implemented in process A”. As shown in FIG. 9, such a record is recorded in the form, and the latest result of the vibration noise problem suppression accuracy of the measure B1 is also utilized in the next model design. Can be prevented in advance.

Also, since the figure is complicated, it is not shown in the figure, but an option of returning one process may be provided. In other words, if it is determined that the expected cost when the current policy is enforced is greater than the sum of the cost due to redesign and the expected cost after the design change, there is an option to return one step. It is best and this embodiment is effective to show it as an objective and rational argument. Specifically, in the current situation where the process has progressed to process B in FIG. 9, it can be seen later that the target achievement probability when implementing the process A1 in process A was 100%. Let us consider a case where the return cost is relatively easy and the return cost is “2”. In this case, the total cost when the process A is returned from the process A to the process B and the measure A1 is further implemented is “4”, and the target achievement probability when the measure A1 is implemented after returning the process is 100%. Since it is known, the expected cost value is “4.0”, which is the strategy with the lowest expected cost value compared to implementing the measure B0 and implementing the measure B1. .

As described above, according to the present embodiment, it is possible to present cost-effectiveness for each option of a measure for suppressing vibration or noise, particularly when designing a newly developed product or considering an improvement plan for an existing product. It is possible to provide a vibration noise countermeasure plan recommendation system.
In addition, according to this embodiment, the problem of the sound source, excitation source and transmission path, the degree of influence of changes in arbitrary elements on the vibration noise of the entire structure, and the probabilistic theory regarding the cost effectiveness of the structure change. Reasoning is possible, and it is possible to provide a judgment material for making a rational judgment for suppressing vibration and noise. Furthermore, as the design / production process progresses, these probabilities are updated as needed based on the obtained data, so the optimal strategy at that time is displayed along with the rationale. For this reason, the cost-effectiveness of each measure is shared as objective information among a plurality of designers, and the decision-making regarding the design policy is facilitated. In addition, design decisions at that time can be traced after the fact, and the decision strategy can be used effectively as past experience knowledge for future designs.

In addition, this invention is not limited to an above-described Example, Various modifications are included.
For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described.

DESCRIPTION OF SYMBOLS 1 ... Vibration noise countermeasure plan recommendation system, 10 ... Product to be measured, 11 ... Site data extraction part, 12 ... Site data storage part, 13 ... Transmission path model data storage part, 14 ... Sound source and transfer characteristic estimation engine, 15 ... Sound source Transfer characteristic probability distribution data storage unit 16 ... laboratory data, 16a ... transfer path model, 16b ... sound source / transfer characteristic probability distribution data, 17 ... laboratory data storage unit, 18 ... contribution rate estimation engine, 19 ... contribution rate Display unit 19a ... Transmission path graphical model display area 19b ... Contribution rate two-dimensional graph display area 19c ... Influence / probability judgment element display area 20 ... User (designer) 30 ... Vibration noise problem factor estimation device, 31 ... Vibration noise transmission path probability reasoning model, 32 ... Vibration noise problem suppression accuracy table, 33 ... Pre- and post-cost list, 34 ... Cost table, 35 ... Gain , 40 ... Vibration noise problem cost effectiveness determination device, 41 ... Cost table generation unit, 42 ... Vibration noise problem suppression accuracy table generation unit, 43 ... Gain table generation unit, 44 ... Input I / F, 45 ... Output I / F 50 ... Game tree (branch diagram of action judgment and situation recognition in the design and production process), 51 ... Design and production process, 52 ... Data extraction work such as inspection, evaluation, and test performed in the design and production process, 53 ... Design and production process 54 ... Action selection in the design production process, 55 ... Situation judgment in the design production process, 56 ... Reflection knowledge in the design production process, 57 ... New discovery solution measure, 59 ... Design data management / storage unit

Claims

Vibration noise problem factor estimating device for estimating vibration noise transmission state of equipment in actual operation state and its contribution rate;
Based on the vibration noise transmission path probability reasoning model generated by the vibration noise problem factor estimation device, a vibration noise problem suppression accuracy table that correlates the presence / absence of implementation of a proactive measure and a post-problem occurrence suppression rate is generated and each input A cost table is generated based on a prior cost list consisting of a prior cost for implementing a measure for suppressing vibration noise problems and a subsequent cost after the occurrence of the problem, and a gain related to expected cost based on the vibration noise problem suppression accuracy table and the cost table Vibration and noise problem cost-effectiveness determination device to generate a table;
A game tree that is a tree structure of action selection and situation determination in each process of design and production of an input product, and a design data management / storage unit that stores the pre-post and post-cost list,
The vibration / noise problem cost-effectiveness determination device generates a gain table regarding expected cost values of each option of action selection in the game tree and presents it to a user.
In the vibration and noise countermeasure plan recommendation system according to claim 1,
A vibration noise countermeasure plan recommendation system characterized in that a gain table relating to an expected cost value of each option of action selection in the game tree and the game tree are displayed side by side on a display unit.
In the vibration and noise countermeasure plan recommendation system according to claim 1,
An anti-vibration measure characterized by having a display unit that highlights a recommended strategy having the smallest cost expected value in the game tree based on a gain table relating to the cost expected value of each option of action selection in the game tree Plan recommendation system.
In the vibration noise countermeasure plan recommendation system according to claim 2 or claim 3,
A vibration noise countermeasure plan recommendation system that displays an expected vibration noise level when an optimum application measure is implemented based on the gain table as a contribution rate two-dimensional graph including variations and fluctuation ranges on the display unit. .
In the vibration noise countermeasure plan recommendation system according to claim 4,
The vibration noise problem cost-effectiveness determination device generates the gain table based on data obtained as a product design and production process proceeds, updates a recommended strategy sequentially, and stores it in the design data management / storage unit. Recommended system for vibration and noise countermeasure plan.
In the vibration noise countermeasure plan recommendation system according to claim 5,
The vibration noise problem cost-effectiveness determination device is based on the vibration noise transmission path probability inference model generated by the vibration noise problem factor estimation device, and the vibration noise problem suppression accuracy that correlates the presence / absence of a prior measure and the subsequent problem occurrence suppression rate. A cost table is generated based on a prior and subsequent cost list composed of a vibration cost problem suppression accuracy table generation unit for generating a table and a prior cost for implementing each vibration noise problem suppression measure and a subsequent cost after the problem occurs. A vibration noise countermeasure plan recommendation system comprising: a cost table generation unit; and a gain table generation unit that generates a gain table based on the vibration noise problem suppression accuracy table and the cost table.
In the vibration noise countermeasure plan recommendation system according to claim 5,
The vibration noise problem factor estimating device is:
At least an on-site data extraction unit that measures on-site data from the evaluation target product,
A transmission path model data storage unit for storing the transmission path model;
Sound source / transfer characteristic estimation for estimating the probability distribution data of the sound source / transfer characteristic corresponding to the element of the evaluation target product from the field data and the transfer path model stored in the transfer path model data storage unit Engine,
Contribution rate estimation for estimating the vibration noise contribution rate based on the transmission route model stored in the transmission route model data storage unit, the sound source / transfer characteristic probability distribution estimated by the sound source / transfer characteristic estimation engine, and the field data And a vibration and noise countermeasure plan recommendation system characterized by comprising an engine.
In the vibration noise countermeasure plan recommendation system according to claim 6,
The vibration noise problem factor estimating device is:
At least an on-site data extraction unit that measures on-site data from the evaluation target product,
A transmission path model data storage unit for storing the transmission path model;
Sound source / transfer characteristic estimation for estimating the probability distribution data of the sound source / transfer characteristic corresponding to the element of the evaluation target product from the field data and the transfer path model stored in the transfer path model data storage unit Engine,
Contribution rate estimation for estimating the vibration noise contribution rate based on the transmission route model stored in the transmission route model data storage unit, the sound source / transfer characteristic probability distribution estimated by the sound source / transfer characteristic estimation engine, and the field data And a vibration and noise countermeasure plan recommendation system characterized by comprising an engine.