BP neural network-based weighing assembly layout method and system

Abstract

Disclosed are a BP neural network-based weighing assembly layout method and system. The method comprises the following steps: step S1, selecting a limited number of drywall weighing products of different specifications as weighing layout samples, and using weighing expectations to perform layout completion analysis on each weighing layout sample to obtain an optimal weighing point for weighing each weighing layout sample; step S2, causing the BP neural network to undergo learning and training on the basis of an attribute feature and the optimal weighing point of each weighing layout sample to obtain a weighing point layout model; and step S3, configuring a weighing support mechanism and a transmitter as a movable structure, the weighing support mechanism and the transmitter performing fixed-point movement according to optimal weighing points of the drywall weighing products outputted by the weighing point layout model. The present invention achieves the optimal layout for the weighing of the drywall weighing products, and the adaptability of weighing relies on modeling of the objective determination of expert experience while ensuring high adaptability and low resource waste.

Description

A weighing component layout method and system based on BP neural network Technical Field

[0001] The present invention relates to the technical field of gypsum board weighing components, and in particular to a weighing component layout method and system based on BP neural network. Background Art

[0002] The average weight of a finished rack of gypsum boards is crucial to their quality. A low weight per unit area can lead to poor bonding, resulting in defective products and negatively impacting production efficiency. The weight per unit area of ​​gypsum board refers to the weight per square meter. Traditionally, several gypsum boards have been stacked together, weighed in batches, and the average value calculated. This inefficient weighing process is no longer suitable for today's automated production.

[0003] Currently, there are many existing devices for quickly weighing gypsum boards. For example, CN202121894810.8 (a continuous weighing device for boards) uses a belt scale to weigh gypsum boards online, in which a belt scale is arranged at a fixed position for continuous transmission and weighing. However, the fixed setting of the weighing components also reflects the shortcomings of existing gypsum board weighing equipment. Due to the upper and lower limits of the weighing, the fixed weighing components need to be specified for the weighing object. Therefore, the number of weighing component combinations is small, resulting in limited product adaptability. Moreover, if the weighing components need to be combined with a higher number of components to improve the weighing adaptability, the fixed combination of a higher number of components will waste hardware resources due to the infrequent occurrence of the occasional weighing object while pursuing adaptability. Existing weighing equipment is equipped with a combination of weighing components during construction. However, most existing weighing component combinations use manual subjective judgment of the weighing object specifications to set the weighing upper and lower limits of the weighing components. That is, the weighing adaptability depends on manual subjective judgment, which makes it difficult to ensure high adaptability and low resource waste at the same time.

[0004] Summary of the Invention

[0005] The purpose of the present invention is to provide a weighing component layout method and system based on BP neural network to solve the technical problem in the prior art that weighing adaptability depends on manual subjective judgment and it is difficult to ensure high adaptability and low resource waste at the same time.

[0006] In order to solve the above technical problems, the present invention specifically provides the following technical solutions:

[0007] A weighing component layout method based on BP neural network includes the following steps:

[0008] Step S1: Select a limited number of gypsum board weighing products of different specifications as weighing layout samples, and perform layout completion analysis on each weighing layout sample using weighing expectation to obtain the optimal weighing point for weighing each weighing layout sample. The weighing expectation is the target result expected to be achieved when weighing each weighing layout sample. The layout completion analysis is to use the weighing expectation to optimally quantify the weighing layout sample to determine the optimal weighing point of the weighing support mechanism and the transmitter when weighing the weighing layout sample. The optimal weighing point is the optimal position of the weighing support mechanism and the transmitter when the weighing expectation optimization target of the weighing layout sample is achieved;

[0009] Step S2: Using a BP neural network to perform learning and training based on the attribute characteristics of each weighing layout sample and the optimal weighing point to obtain a weighing point layout model;

[0010] Step S3: The weighing support mechanism and the transmitter are configured as movable structures. The weighing support mechanism and the transmitter are moved to a fixed point according to the optimal weighing point of the gypsum board weighing product output by the weighing point layout model, so as to achieve an optimized layout for weighing the gypsum board weighing product.

[0011] Among them, a weighing sensor is provided on the upper surface of the weighing support mechanism, and the weighing support mechanism is used to support the gypsum board weighing product and the weighing sensor to complete the weighing of the gypsum board weighing product to obtain a digital signal representing the weight of the gypsum board weighing product. The transmitter is communicatively connected with the weighing sensor, and the transmitter is used to transmit the digital signal representing the weight of the gypsum board weighing product to the weighing host computer to display the weighing weight of the gypsum board weighing product.

[0012] As a preferred solution of the present invention, the method of performing layout completeness analysis on each weighing layout sample using weighing expectation to obtain the optimal weighing point for weighing each weighing layout sample includes:

[0013] Construct the expected weighing accuracy, the quantitative expression of which is:

[0014] Where K is the expected weighing accuracy, G real is the true value of the weight of the weighing layout sample, P i is the i-th weighing point of the weighing support mechanism, G c (P i ) is the weighing weight of the i-th weighing point of the weighing support mechanism, is the weighing weight of all weighing points of the weighing support mechanism and / or the weighing weight of the weighing layout sample, n is the number of weighing points of the weighing support mechanism, and i is a counting variable;

[0015] Construct a weighed cost expectation, the quantitative expression of which is:

[0016] Where L is the expected cost of weighing, M c is the unit cost of the weighing support mechanism, M z is the unit cost of the weighing sensor, M l is the unit cost of the transmitter, n is the number of weighing points of the weighing support mechanism, m is the number of weighing points of the transmitter, m:n=D, D is the proportional constant;

[0017] Maximize the expected weighing accuracy maxK and minimize the expected weighing cost minL, and jointly solve the maximized weighing accuracy expectation and the minimized weighing cost expectation to obtain the number of weighing points and the optimal weighing point of the weighing support mechanism in the weighing layout sample. max is the maximization operator and min is the minimization operator.

[0018] Based on the adaptability rule of the number of weighing support mechanisms and transmitters m:n=D, the number of weighing points of the transmitters in the weighing layout sample is obtained according to the adaptability of the number of weighing points of the weighing support mechanism in the weighing layout sample;

[0019] Based on the position adaptability rule of the weighing support mechanism and the transmitter, the optimal weighing point of the transmitter in the weighing layout sample is obtained according to the optimal weighing point adaptability of the weighing support mechanism in the weighing layout sample.

[0020] As a preferred solution of the present invention, the BP neural network is used to learn and train the weighing point layout model based on the attribute characteristics of each weighing layout sample and the optimal weighing point, including:

[0021] The attribute characteristics of the weighing layout sample are used as the input of the BP neural network, and the optimal weighing point of the weighing layout sample is used as the output of the BP neural network;

[0022] Using a BP neural network to learn and train the input items and output items of the BP neural network to obtain the weighing point layout model;

[0023] The model expression of the weighing point location layout model is: [P, Q] = BP (S);

[0024] Where P is the optimal weighing point of the weighing support mechanism, Q is the optimal weighing point of the transmitter, S is the attribute feature, and BP is the neural network.

[0025] As a preferred solution of the present invention, the attribute characteristics include length specifications, shape specifications and material specifications of the gypsum board weighing product.

[0026] As a preferred solution of the present invention, the unit cost includes unit installation cost, unit device cost and unit operation cost.

[0027] As a preferred solution of the present invention, during the layout completeness analysis process, the weighing weight of the weighing point is normalized with the unit cost of the weighing support mechanism, the unit cost of the weight sensor, and the unit cost of the transmitter.

[0028] As a preferred solution of the present invention, a multi-objective optimization algorithm is used to jointly solve the maximization of weighing accuracy expectation and the minimization of weighing cost expectation.

[0029] As a preferred embodiment of the present invention, the present invention provides a layout system for implementing the weighing component layout method, comprising a weighing support mechanism, a weighing sensor, a transmitter, a gypsum board property acquisition device, a host computer, and a cylinder control system, wherein;

[0030] The weighing support mechanism is used to support the gypsum board weighing product to contact the weighing sensor in a press-type manner, so that the weighing sensor obtains a digital signal representing the weight of the gypsum board weighing product;

[0031] The weighing sensor is used to obtain a digital signal representing the weight of the gypsum board weighing product;

[0032] The transmitter is used to transmit a digital signal representing the weight of the gypsum board weighing product to the weighing host computer;

[0033] The gypsum board property collection device is used to collect the property characteristics of the gypsum board weighing product;

[0034] The host computer is used to obtain the optimal weighing point of the gypsum board weighing product based on the attribute characteristics of the gypsum board weighing product fed back by the gypsum board attribute collection device using the weighing point layout model, and display the weighing weight of the gypsum board weighing product;

[0035] The cylinder control system is used to control the weighing support mechanism to move to the optimal weighing point according to the optimal weighing point of the gypsum board weighing product output by the host computer.

[0036] As a preferred solution of the present invention, the weighing sensor is arranged on the upper surface of the weighing support mechanism and is communicatively connected to the transmitter, the transmitter is communicatively connected to the host computer, the weighing support mechanism is communicatively connected to the cylinder control system, the cylinder control system is communicatively connected to the host computer, and the gypsum board property acquisition device is communicatively connected to the host computer. A weighing support mechanism is provided at the optimal weighing point of each weighing support mechanism, and two weighing sensors are provided on each weighing support mechanism.

[0037] As a preferred solution of the present invention, the construction of the weighing point layout model includes:

[0038] Construct the expected weighing accuracy, the quantitative expression of which is:

[0039] Where K is the expected weighing accuracy, G real is the true value of the weight of the weighing layout sample, P i is the i-th weighing point of the weighing support mechanism, G c (P i ) is the weighing weight of the i-th weighing point of the weighing support mechanism, is the weighing weight of all weighing points of the weighing support mechanism and / or the weighing weight of the weighing layout sample, n is the number of weighing points of the weighing support mechanism, and i is a counting variable;

[0040] Construct a weighed cost expectation, the quantitative expression of which is:

[0041] Where L is the expected cost of weighing, M c is the unit cost of the weighing support mechanism, M z is the unit cost of the weighing sensor, M l is the unit cost of the transmitter, n is the number of weighing points of the weighing support mechanism, m is the number of weighing points of the transmitter, m:n=D, D is the proportional constant;

[0042] Maximize the expected weighing accuracy maxK and minimize the expected weighing cost minL, and jointly solve the maximized weighing accuracy expectation and the minimized weighing cost expectation to obtain the number of weighing points and the optimal weighing point of the weighing support mechanism in the weighing layout sample. max is the maximization operator and min is the minimization operator.

[0043] Based on the adaptability rule of the number of weighing support mechanisms and transmitters m:n=D, the number of weighing points of the transmitters in the weighing layout sample is obtained according to the adaptability of the number of weighing points of the weighing support mechanism in the weighing layout sample;

[0044] Based on the adaptability rule of the weighing support mechanism and the transmitter, the optimal weighing point of the transmitter in the weighing layout sample is obtained according to the optimal weighing point adaptability of the weighing support mechanism in the weighing layout sample.

[0045] The attribute characteristics of the weighing layout sample are used as the input of the BP neural network, and the optimal weighing point of the weighing layout sample is used as the output of the BP neural network;

[0046] Using a BP neural network to learn and train the input items and output items of the BP neural network to obtain the weighing point layout model;

[0047] The model expression of the weighing point location layout model is: [P, Q] = BP (S);

[0048] Where P is the optimal weighing point of the weighing support mechanism, Q is the optimal weighing point of the transmitter, S is the attribute feature, and BP is the neural network.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] The present invention utilizes a BP neural network to perform learning and training based on the attribute characteristics of each weighing layout sample and the optimal weighing point to obtain a weighing point layout model, and sets the weighing support mechanism and the transmitter as a movable structure. The weighing support mechanism and the transmitter perform fixed-point movement according to the optimal weighing point of the gypsum board weighing product output by the weighing point layout model, so as to achieve the optimal layout of the weighing of the gypsum board weighing product, and objectively judge the expert experience of the weighing adaptability dependency model while ensuring high adaptability and low resource waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other implementation drawings based on the provided drawings without inventive effort.

[0052] FIG1 is a flow chart of a weighing component layout method according to an embodiment of the present invention;

[0053] FIG2 is a structural block diagram of a layout system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] The fixed setting of the weighing components also reflects the defects of the existing gypsum board weighing equipment. Due to the upper and lower limits of the weighing, the fixed weighing components need to be specified for the weighing object. Therefore, the number of weighing component combinations is small, which will lead to limited product adaptability. Moreover, if the weighing components need to select a higher number of component combinations in order to improve the weighing adaptability, the fixed combination of higher number of components will waste hardware resources due to the infrequent appearance of occasional weighing objects while pursuing adaptability. The existing weighing equipment has a combination setting of weighing components when it is built, but most of the existing weighing component combinations use manual subjective judgment of the weighing object specifications to set the weighing upper and lower limits of the weighing components, that is, the weighing adaptability depends on manual subjective judgment, which is difficult to ensure high adaptability and low resource waste at the same time. Therefore, the present invention provides a weighing component layout method, constructs a weighing point layout model, realizes the optimal layout of the weighing of gypsum board weighing products, and the weighing adaptability depends on the objective judgment of expert experience modeling, while ensuring high adaptability and low resource waste.

[0056] As shown in FIG1 , the present invention provides a weighing component layout method based on a BP neural network, comprising the following steps:

[0057] Step S1: Select a limited number of gypsum board weighing products of different specifications as weighing layout samples, and perform layout completion analysis on each weighing layout sample using weighing expectation to obtain the optimal weighing point for weighing each weighing layout sample. The weighing expectation is the target result expected to be achieved when weighing each weighing layout sample. The layout completion analysis is to use the weighing expectation to optimally quantify the weighing layout sample to determine the optimal weighing point of the weighing support mechanism and the transmitter when weighing the weighing layout sample. The optimal weighing point is the optimal position of the weighing support mechanism and the transmitter when the weighing expectation optimization target of the weighing layout sample is achieved.

[0058] Step S2: Using a BP neural network to perform learning and training based on the attribute characteristics of each weighing layout sample and the optimal weighing point to obtain a weighing point layout model;

[0059] Step S3: The weighing support mechanism and the transmitter are configured as movable structures. The weighing support mechanism and the transmitter are moved to a fixed point according to the optimal weighing point of the gypsum board weighing product output by the weighing point layout model, so as to achieve an optimized layout for weighing the gypsum board weighing product.

[0060] Among them, a weighing sensor is provided on the upper surface of the weighing support mechanism, and the weighing support mechanism is used to support the gypsum board weighing product and the weighing sensor to press-contact to complete the weighing of the gypsum board weighing product, so as to obtain a digital signal representing the weight of the gypsum board weighing product. The transmitter is communicated with the weighing sensor, and the transmitter is used to transmit the digital signal representing the weight of the gypsum board weighing product to the weighing host computer to display the weighing weight of the gypsum board weighing product.

[0061] The present invention uses layout completeness analysis to determine the optimal weighing points of the weighing support mechanism and the transmitter, and can realize the optimal weighing position setting of the weighing component according to the product specifications of the gypsum board weighing product, avoid the waste of hardware resources, and ensure the accuracy and economy of weighing. At the same time, the BP neural network is used to construct a weighing point layout model, which can enable the weighing point layout model to integrate the advantages of layout completeness analysis, so that the weighing points of various gypsum board weighing products can be calculated by modeling. The weighing points can accurately weigh various gypsum board weighing products and maintain the minimum weighing component device, installation and operation costs. The weighing support mechanism and the transmitter are set as movable structures to match the weighing point layout model to achieve data and hardware coordination. The optimal layout scheme calculated by the model is implemented for various gypsum board weighing products. The weighing component hardware is implemented, the weighing adaptability of the weighing combination layout is improved, and the flexible adjustment of the component layout avoids the waste of hardware resources caused by the fixed component layout.

[0062] The present invention uses layout completeness analysis to analyze the optimal layout of weighing components. We quantify the weighing accuracy expectation and weighing cost expectation as two layout expectations. Through quantitative analysis, we find the optimal layout of weighing components that can balance weighing accuracy and weighing cost. That is, the optimal layout can maintain low weighing cost while ensuring high weighing accuracy. The details are as follows:

[0063] The layout completion analysis of each weighing layout sample is performed using weighing expectation to obtain the optimal weighing point for each weighing layout sample, including:

[0064] Construct the expected weighing accuracy. The quantitative expression of the expected weighing accuracy is:

[0065] Where K is the expected weighing accuracy, G real is the true value of the weight of the weighing layout sample, P i is the i-th weighing point of the weighing support mechanism, G c (P i ) is the weighing weight of the i-th weighing point of the weighing support mechanism, is the weighing weight of all weighing points of the weighing support mechanism and / or the weighing weight of the weighing layout sample, n is the number of weighing points of the weighing support mechanism, and i is a counting variable;

[0066] Construct the weighed cost expectation, and the quantitative expression of the weighed cost expectation is:

[0067] Where L is the expected cost of weighing, M c is the unit cost of the weighing support mechanism, M z is the unit cost of the weighing sensor, M l is the unit cost of the transmitter, n is the number of weighing points of the weighing support mechanism, m is the number of weighing points of the transmitter, m:n=D, D is the proportional constant;

[0068] Maximize the expected weighing accuracy maxK and minimize the expected weighing cost minL, and jointly solve the maximized weighing accuracy expectation and the minimized weighing cost expectation to obtain the number of weighing points and the optimal weighing point of the weighing support mechanism in the weighing layout sample. max is the maximization operator and min is the minimization operator.

[0069] Based on the adaptability rule of the number of weighing support mechanisms and transmitters, m:n=D, for example, 12 load cells are adapted to 2 transmitters; 14 load cells are adapted to 3 transmitters; 16 load cells are adapted to 4 transmitters; 18 load cells are adapted to 5 transmitters; and 20 load cells are adapted to 6 transmitters, the number of weighing points of the transmitters in the weighing layout sample is obtained according to the adaptability of the number of weighing points of the weighing support mechanism in the weighing layout sample.

[0070] Based on the position adaptability rule of the weighing support mechanism and the transmitter, the optimal weighing point of the transmitter in the weighing layout sample is obtained according to the optimal weighing point adaptability of the weighing support mechanism in the weighing layout sample.

[0071] The present invention uses layout completeness analysis to analyze the optimal layout of weighing components to obtain model training samples, and uses BP neural network to learn and train attribute characteristics and optimal weighing points, so as to directly obtain the optimal weighing points based on the attribute characteristics of gypsum board weighing products. Since the training samples come from layout completeness analysis, the weighing point layout model obtained by training has the advantages of high weighing accuracy and low weighing cost, and realizes the objective judgment of modeled weighing adaptability dependence and modeled expert experience. The model construction is specifically as follows:

[0072] The BP neural network is used to learn and train the weighing point layout model based on the attribute characteristics of each weighing layout sample and the optimal weighing point, including:

[0073] The attribute characteristics of the weighing layout sample are used as the input of the BP neural network, and the optimal weighing point of the weighing layout sample is used as the output of the BP neural network;

[0074] The BP neural network is used to train the input items and output items of the BP neural network to obtain the weighing point layout model;

[0075] The model expression of the weighing point layout model is: [P,Q]=BP(S);

[0076] Where P is the optimal weighing point of the weighing support mechanism, Q is the optimal weighing point of the transmitter, S is the attribute feature, and BP is the neural network.

[0077] The attribute characteristics include the length specifications, shape specifications and material specifications of the gypsum board weighing products.

[0078] Unit cost includes unit installation cost, unit device cost and unit operating cost.

[0079] During the layout completeness analysis, the weighing weight at the weighing point is normalized with the unit cost of the weighing support mechanism, the unit cost of the weight sensor, and the unit cost of the transmitter.

[0080] The multi-objective optimization algorithm is used to jointly solve the problem of maximizing the expected weighing accuracy and minimizing the expected weighing cost.

[0081] The present invention utilizes a layout system to set the weighing support mechanism and transmitter as a movable structure to match the weighing point layout model, realize the coordination of data and hardware, and implement the weighing component hardware implementation based on the optimized layout scheme calculated by the model for various gypsum board weighing products, thereby improving the weighing adaptability of the weighing combination layout and flexibly adjusting the component layout to avoid the waste of hardware resources caused by the fixed component layout. The system settings are as follows:

[0082] As shown in FIG2 , a layout system for implementing a weighing component layout method includes a weighing support mechanism, a weighing sensor, a transmitter, a gypsum board property acquisition device, a host computer, and a cylinder control system, wherein;

[0083] The weighing support mechanism is used to support the gypsum board weighing product to contact the weighing sensor in a press-type manner, so that the weighing sensor can obtain a digital signal representing the weight of the gypsum board weighing product;

[0084] The load cell is used to obtain a digital signal representing the weight of the gypsum board weighing product;

[0085] The transmitter is used to transmit the digital signal representing the weight of the gypsum board weighing product to the weighing host computer;

[0086] The gypsum board property collection device is used to collect the property characteristics of the gypsum board weighing product;

[0087] The upper computer is used to obtain the optimal weighing point of the gypsum board weighing product based on the attribute characteristics of the gypsum board weighing product fed back by the gypsum board attribute collection device by using the weighing point layout model, and to display the weighing weight of the gypsum board weighing product;

[0088] The cylinder control system is used to control the weighing support mechanism to move to the optimal weighing point according to the optimal weighing point of the gypsum board weighing product output by the upper computer.

[0089] The weighing sensor is arranged on the upper surface of the weighing support mechanism and is communicated with the transmitter, the transmitter is communicated with the host computer, the weighing support mechanism is communicated with the cylinder control system, the cylinder control system is communicated with the host computer, and the gypsum board property collection device is communicated with the host computer. A weighing support mechanism is provided at the optimal weighing point of each weighing support mechanism, and two weighing sensors are provided on each weighing support mechanism.

[0090] The construction of the weighing point layout model includes:

[0091] Construct the expected weighing accuracy. The quantitative expression of the expected weighing accuracy is:

[0092] Where K is the expected weighing accuracy, G real is the true value of the weight of the weighing layout sample, P i is the i-th weighing point of the weighing support mechanism, G c (P i ) is the weighing weight of the i-th weighing point of the weighing support mechanism, is the weighing weight of all weighing points of the weighing support mechanism and / or the weighing weight of the weighing layout sample, n is the number of weighing points of the weighing support mechanism, and i is a counting variable;

[0093] Construct the weighed cost expectation, and the quantitative expression of the weighed cost expectation is:

[0094] Where L is the expected cost of weighing, M c is the unit cost of the weighing support mechanism, M z is the unit cost of the weighing sensor, M l is the unit cost of the transmitter, n is the number of weighing points of the weighing support mechanism, m is the number of weighing points of the transmitter, m:n=D, D is the proportional constant;

[0095] Maximize the expected weighing accuracy maxK and minimize the expected weighing cost minL, and jointly solve the maximized weighing accuracy expectation and the minimized weighing cost expectation to obtain the number of weighing points and the optimal weighing point of the weighing support mechanism in the weighing layout sample. max is the maximization operator and min is the minimization operator.

[0096] Based on the adaptability rule of the number of weighing support mechanisms and transmitters m:n=D, the number of weighing points of the transmitters in the weighing layout sample is obtained according to the adaptability of the number of weighing points of the weighing support mechanism in the weighing layout sample;

[0097] Based on the adaptability rule of the weighing support mechanism and the transmitter, the optimal weighing point of the transmitter in the weighing layout sample is obtained according to the optimal weighing point adaptability of the weighing support mechanism in the weighing layout sample.

[0098] The attribute characteristics of the weighing layout sample are used as the input of the BP neural network, and the optimal weighing point of the weighing layout sample is used as the output of the BP neural network;

[0099] The BP neural network is used to train the input items and output items of the BP neural network to obtain the weighing point layout model;

[0100] The model expression of the weighing point layout model is: [P,Q]=BP(S);

[0101] Where P is the optimal weighing point of the weighing support mechanism, Q is the optimal weighing point of the transmitter, S is the attribute feature, and BP is the neural network.

[0102] The present invention uses layout completeness analysis to determine the optimal weighing points of the weighing support mechanism and the transmitter, and can realize the optimal weighing position setting of the weighing component according to the product specifications of the gypsum board weighing product, avoid the waste of hardware resources, and ensure the accuracy and economy of weighing. At the same time, the BP neural network is used to construct a weighing point layout model, which can enable the weighing point layout model to integrate the advantages of layout completeness analysis, so that the weighing points of various gypsum board weighing products can be calculated by modeling. The weighing points can accurately weigh various gypsum board weighing products and maintain the minimum weighing component device, installation and operation costs. The weighing support mechanism and the transmitter are set as movable structures to match the weighing point layout model to achieve data and hardware coordination. The optimal layout scheme calculated by the model is implemented for various gypsum board weighing products. The weighing component hardware is implemented, the weighing adaptability of the weighing combination layout is improved, and the flexible adjustment of the component layout avoids the waste of hardware resources caused by the fixed component layout.

[0103] The present invention uses layout completeness analysis to analyze the optimal layout of weighing components. It quantitatively analyzes the weighing accuracy expectation and weighing cost expectation as two layout expectations. Through quantitative analysis, the optimal layout of weighing components that can balance weighing accuracy and weighing cost is obtained. That is, the optimal layout can maintain low weighing cost while ensuring high weighing accuracy.

[0104] The present invention uses layout completeness analysis to perform optimal layout analysis of weighing components to obtain model training samples, and uses BP neural network to learn and train attribute characteristics and optimal weighing points, so as to obtain the optimal weighing points directly through the attribute characteristics of gypsum board weighing products. Since the training samples come from layout completeness analysis, the weighing point layout model obtained by training has the advantages of high weighing accuracy and low weighing cost, and realizes the objective judgment of modeled weighing adaptability dependence and modeled expert experience.

[0105] The present invention utilizes a BP neural network to perform learning and training based on the attribute characteristics of each weighing layout sample and the optimal weighing point to obtain a weighing point layout model, and sets the weighing support mechanism and the transmitter as a movable structure. The weighing support mechanism and the transmitter perform fixed-point movement according to the optimal weighing point of the gypsum board weighing product output by the weighing point layout model, so as to achieve the optimal layout of the weighing of the gypsum board weighing product, and objectively judge the expert experience of the weighing adaptability dependency model while ensuring high adaptability and low resource waste.

[0106] The above embodiments are merely exemplary embodiments of the present application and are not intended to limit the scope of the present application. The scope of protection of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and scope of protection of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the scope of protection of the present application.

Claims

1. A weighing component layout method based on BP neural network, characterized in that: The following steps are involved: Step S1, select a limited amount of gypsum board weighing products of different specifications as weighing layout samples, and use the weighing expectation to perform layout completion analysis on each weighing layout sample to obtain the optimal weighing point for weighing each weighing layout sample, wherein the weighing expectation is the target result expected to be obtained when weighing each weighing layout sample, and the layout completion analysis is to use the weighing expectation to perform optimal quantification on the weighing layout sample to determine the optimal weighing point of the weighing support mechanism and the transmitter when weighing the weighing layout sample, and the optimal weighing point is the best position of the weighing support mechanism and the transmitter when the weighing expectation optimization target of the weighing layout sample is completed; Step S2, using BP neural network to perform learning and training based on the attribute characteristics of each weighing layout sample and the optimal weighing point to obtain a weighing point layout model; Step S3, setting the weighing support mechanism and the transmitter as a movable structure, and moving the weighing support mechanism and the transmitter at a fixed point according to the optimal weighing point of the gypsum board weighing product output by the weighing point layout model, so as to achieve the optimal layout of weighing the gypsum board weighing product; Among them, a weighing sensor is arranged on the upper surface of the weighing support mechanism, and the weighing support mechanism is used to support the gypsum board weighing product and the weighing sensor to press-contact to complete the weighing of the gypsum board weighing product, so as to obtain a digital signal representing the weight of the gypsum board weighing product, and the transmitter is communicatively connected with the weighing sensor, and the transmitter is used to transmit the digital signal representing the weight of the gypsum board weighing product to the weighing host computer to display the weighing weight of the gypsum board weighing product.

2. The weighing component layout method based on BP neural network according to claim 1 is characterized in that: The method of performing layout completion analysis on each weighing layout sample using weighing expectation to obtain the optimal weighing point for weighing each weighing layout sample comprises: Construct the expected weighing accuracy, the quantitative expression of which is: Where K is the expected weighing accuracy, G real is the true value of the weight of the weighing layout sample, P i is the i-th weighing point of the weighing support mechanism, G c (P i ) is the weighing weight of the ith weighing point of the weighing support mechanism, Weighing weight and / or weighing layout of all weighing points of the weighing support mechanism The weighing weight of the sample, n is the number of weighing points of the weighing support mechanism, and i is the counting variable; Construct a weighed cost expectation, the quantitative expression of which is: In the formula, L is the expected cost of weighing, M c is the unit cost of the weighing support mechanism, M z is the unit cost of the weighing sensor, M l is the unit cost of the transmitter, n is the number of weighing points of the weighing support mechanism, m is the number of weighing points of the transmitter, m:n=D, D is the proportional constant; Maximize the expected weighing accuracy maxK and minimize the expected weighing cost minL, and jointly solve the maximized weighing accuracy expectation and the minimized weighing cost expectation to obtain the number of weighing points and the optimal weighing points of the weighing support mechanism in the weighing layout sample, where max is the maximization operator and min is the minimization operator; Based on the number adaptability rule of the weighing support mechanism and the transmitter m:n=D, the number of weighing points of the transmitter in the weighing layout sample is obtained according to the adaptability of the number of weighing points of the weighing support mechanism in the weighing layout sample; Based on the position adaptability rules of the weighing support mechanism and the transmitter, the optimal weighing point position of the transmitter in the weighing layout sample is obtained according to the optimal weighing point position adaptability of the weighing support mechanism in the weighing layout sample.

3. The weighing component layout method based on BP neural network according to claim 2 is characterized in that: The method of using the BP neural network to learn and train based on the attribute characteristics of each weighing layout sample and the optimal weighing point to obtain a weighing point layout model includes: The attribute characteristics of the weighing layout samples are used as the input items of the BP neural network, and the optimal weighing point of the weighing layout samples is used as the output item of the BP neural network; Using a BP neural network to learn and train the input items and output items of the BP neural network to obtain the weighing point location layout model; The model expression of the weighing point location layout model is: [P,Q]=BP(S); Where P is the optimal weighing point of the weighing support mechanism, Q is the optimal weighing point of the transmitter, S is the attribute feature and BP is the neural network.

4. The weighing component layout method based on BP neural network according to claim 1 is characterized in that: The attribute characteristics include length specifications, shape specifications and material specifications of the gypsum board weighing product.

5. The weighing component layout method based on BP neural network according to claim 2 is characterized in that: The unit cost includes unit installation cost, unit device cost and unit operation cost.

6. The weighing component layout method based on BP neural network according to claim 2 is characterized in that: In the layout completeness analysis process, the weighing weight of the weighing point is normalized with the unit cost of the weighing support mechanism, the unit cost of the weighing sensor, and the unit cost of the transmitter.

7. A weighing component layout method based on BP neural network according to claim 2, characterized in that: The multi-objective optimization algorithm is used to jointly solve the maximization of the expected weighing accuracy and the minimization of the expected weighing cost.

8. A layout system for implementing the weighing component layout method according to any one of claims 1 to 7, characterized in that: It includes a weighing support mechanism, a weighing sensor, a transmitter, a gypsum board property collection device, a host computer, and a cylinder control system, among which; The weighing support mechanism is used to support the gypsum board weighing product to be in pressure contact with the weighing sensor, so that the weighing sensor obtains a digital signal representing the weight of the gypsum board weighing product; The weighing sensor is used to obtain a digital signal representing the weight of the gypsum board weighing product; The transmitter is used to transmit a digital signal representing the weight of the gypsum board weighing product to the weighing host computer; The gypsum board property collection device is used to collect the property characteristics of the gypsum board weighing product; The host computer is used to obtain the optimal weighing point of the gypsum board weighing product based on the property characteristics of the gypsum board weighing product fed back by the gypsum board property collection device by using the weighing point layout model, and display the weighing weight of the gypsum board weighing product; The cylinder control system is used to control the weighing support mechanism to move to the optimal weighing point according to the optimal weighing point of the gypsum board weighing product output by the upper computer.

9. A layout system according to claim 8, characterized in that: The weighing sensor is arranged on the upper surface of the weighing support mechanism and is communicatively connected to the transmitter, the transmitter is communicatively connected to the host computer, the weighing support mechanism is communicatively connected to the cylinder control system, the cylinder control system is communicatively connected to the host computer, the gypsum board property acquisition device is communicatively connected to the host computer, a weighing support mechanism is arranged at the optimal weighing point of each weighing support mechanism, and two weighing sensors are arranged on each weighing support mechanism.

10. A layout system according to claim 8, characterized in that: The construction of the weighing point location layout model includes: Construct the expected weighing accuracy, the quantitative expression of which is: Where K is the expected weighing accuracy, G real is the true value of the weight of the weighing layout sample, P i is the i-th weighing point of the weighing support mechanism, G c (P i ) is the weighing weight of the ith weighing point of the weighing support mechanism, is the weighing weight of all weighing points of the weighing support mechanism and / or the weighing weight of the weighing layout sample, n is the number of weighing points of the weighing support mechanism, and i is a counting variable; Construct a weighed cost expectation, the quantitative expression of which is: In the formula, L is the expected cost of weighing, M c is the unit cost of the weighing support mechanism, M z is the unit cost of the weighing sensor, M l is the unit cost of the transmitter, n is the number of weighing points of the weighing support mechanism, m is the number of weighing points of the transmitter, m:n=D, D is the proportional constant; Maximize the expected weighing accuracy maxK and minimize the expected weighing cost minL, and jointly solve the maximized weighing accuracy expectation and the minimized weighing cost expectation to obtain the number of weighing points and the optimal weighing points of the weighing support mechanism in the weighing layout sample, where max is the maximization operator and min is the minimization operator; Based on the number adaptability rule of the weighing support mechanism and the transmitter m:n=D, the number of weighing points of the transmitter in the weighing layout sample is obtained according to the adaptability of the number of weighing points of the weighing support mechanism in the weighing layout sample; Based on the adaptability rules of the weighing support mechanism and the transmitter, the optimal weighing point of the transmitter in the weighing layout sample is obtained according to the optimal weighing point adaptability of the weighing support mechanism in the weighing layout sample. The attribute characteristics of the weighing layout samples are used as the input items of the BP neural network, and the optimal weighing point of the weighing layout samples is used as the output item of the BP neural network; Using a BP neural network to learn and train the input items and output items of the BP neural network to obtain the weighing point location layout model; The model expression of the weighing point location layout model is: [P,Q]=BP(S); Where P is the optimal weighing point of the weighing support mechanism, Q is the optimal weighing point of the transmitter, S is the attribute feature, and BP is the neural network.

Images