Winch, winch system, heavy object mass estimation device and heavy object mass estimation method

Abstract

The present invention can estimate the mass of the weight of the hoist with high accuracy. The hoist of the present invention, which can move the suspended weight in the vertical direction, includes: an operation information acquisition unit for acquiring the operation information of the hoist; and a weight mass estimation unit for estimating the mass of the weight of the hoist from the operation information acquired by the operation information acquisition unit, the weight mass estimation unit estimates the weight based on a weight mass estimation formula, wherein the weight mass estimation formula is pre-learned using data that associates the operation information with the weight. Thus, it is possible to estimate the mass of the weight that can reflect the actions of the real world that cannot be fully modeled by the model-based method, and thus it is possible to estimate the mass of the weight of the hoist with high accuracy.

Description

Technical Field

[0001] The present invention relates to a winch. Background Art

[0002] In recent years, with the aging of skilled operators of winches and the shortage of manpower caused by the increase in the number of winches, the number of unskilled operators with little experience is increasing in the winch industry. Under such circumstances, a winch that can also be used by unskilled operators to achieve safe and efficient transportation is required. In a winch, it is required to be able to detect the weight of a heavy object not only to inform the operator of the weight of the heavy object and assist the transportation operation, but also to prevent overload, predict the service life, and detect the weight off the ground.

[0003] In order to detect the mass of a heavy object, for example, a measuring sensor such as a load cell can be added, but when this method is used, the cost for adding the sensor increases. Therefore, a method for estimating the mass of a heavy object without adding a sensor is required.

[0004] In order to realize such a requirement, there is, for example, Patent Document 1. Patent Document 1 discloses a load detection method for a winch, in which "the output torque estimation unit 151 obtains the current value input to the motor 13 from the power converter 14, and calculates the estimated value of the output torque of the motor 13, i.e., the output torque estimated value, based on the obtained current value. The load estimation unit 152 estimates the load value of the heavy object 17 based on the output torque estimated value calculated by the output torque estimation unit 151, the reduction ratio of the reducer 12, the effective radius of the winch drum, and the number of ropes set by the rope number setting unit 153" (reference abstract).

[0005] Prior art literature

[0006] Patent Literature

[0007] Patent Document 1: Japanese Patent Application Publication No. 2018-2477 Summary of the invention

[0008] Problems to be solved by the invention

[0009] In the technology described in Patent Document 1, the output torque is calculated from the current value of the motor based on the motor model, and the mass of the heavy object is estimated from the output torque based on the winch model. However, such a model-based estimation method has the problem of poor estimation accuracy due to model errors caused by the inability to fully model the actions of the real world. For example, in the motor model, it is difficult to correctly estimate the output torque based on the current value of the motor because it cannot detect the friction between the bearings and the reducer, and the influence of the error caused by the cogging torque generated by the rotation of the motor even in the non-powered state.

[0010] The present invention has been made in view of the above background, and an object of the present invention is to estimate the mass of a load of a hoist with high accuracy.

[0011] Technical solutions to solve problems

[0012] In order to solve the above-mentioned problems, the present invention is a winch capable of moving a suspended heavy object in a vertical direction, comprising: an operation information acquisition unit for acquiring operation information of the winch; and a weight mass estimation unit for estimating the weight of the weight of the winch from the operation information acquired by the operation information acquisition unit, the weight mass estimation unit estimating the weight of the weight based on a weight mass estimation formula, wherein the weight mass estimation formula is pre-learned using data that associates the operation information with the weight mass.

[0013] Effects of the Invention

[0014] According to the present invention, it is possible to estimate the mass of a heavy object that can reflect the behavior of the real world that cannot be completely modeled by a model-based method, and thus it is possible to estimate the mass of the heavy object of the hoist with high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a figure which shows the structure of the crane using the hoisting machine of this invention.

[0016] Figure 2 It is a diagram showing a configuration example of a hoisting machine and an operation terminal according to the first embodiment.

[0017] Figure 3 This is a schematic diagram showing that a heavy object is lifted up by a winch and is lifted off the ground.

[0018] Figure 4 It is a schematic diagram showing the positional relationship of the heavy object relative to the cart when it is off the ground.

[0019] Figure 5 This is a flowchart showing the flow of the overall processing performed in the hoisting machine according to the first embodiment.

[0020] Figure 6 This is a flowchart showing the detailed flow of estimating the weight mass during operation.

[0021] Figure 7 It is a diagram showing a configuration example of a hoisting machine and a heavy object mass estimation device according to a second embodiment.

[0022] Figure 8 1 is a flowchart showing the flow of the overall processing performed in the heavy object mass estimating device according to the second embodiment.

[0023] Fig. 9It is a diagram showing a configuration example of a hoisting machine, an operation terminal, and a learning device according to a third embodiment.

[0024] Fig.10 This is a flowchart showing the flow of the overall processing performed in the hoisting machine system according to the third embodiment.

[0025] Fig.11 This is a flowchart showing the detailed flow of data collection of the hoist in the lift-off test.

[0026] Fig.12 This is a flowchart showing the detailed flow of learning the heavy object mass estimation formula.

[0027] Fig.13 This is a diagram showing an example of a system including a plurality of hoists.

[0028] Fig.14 This is a diagram showing an example of a system in which a learning device exists on a cloud server. DETAILED DESCRIPTION

[0029] Next, a mode for carrying out the present invention (referred to as “embodiment”) will be described in detail with reference to the drawings as appropriate.

[0030] [First embodiment]

[0031] refer to Figure 1 to Figure 6 , the first embodiment is described.

[0032] Figure 1 This is a diagram showing the structure of a crane using a winch according to the present embodiment. The crane 500 is composed of a track 501 provided along the walls on both sides of a building (not shown), a bridge 502 moving on the track 501, and a trolley 503 moving along the bridge 502. A winch 1 is provided on the trolley 503, and is used to raise or lower a rope, chain, etc., from which a weight 504 is suspended, thereby raising or lowering the weight 504. In the crane 500, motors for driving the bridge 502, the trolley 503, and the winch 1 are provided, respectively, and are driven according to an operation instruction from an operation terminal 2. It should be noted that the weight 504 is an object to be transported by the crane, and is not a component of the crane.

[0033] Figure 2 1 is a block diagram of a system related to a hoisting machine according to the present embodiment. The hoisting machine system is composed of a hoisting machine 1 and an operation terminal 2 .

[0034] The hoist 1 has a motor 101, a motor operation information acquisition unit 131, a motor control unit 132, and a weight mass estimation unit 133. In addition, the hoist 1 has hardware such as a CPU and other computing devices, a main storage device such as a semiconductor memory, an auxiliary storage device such as a hard disk, and a communication device. Then, the operation command input by the operation terminal 2 is received via the communication device, and the computing device executes the program stored in the main storage device, thereby realizing the various functions of the motor operation information acquisition unit 131 and the motor control unit 132. The details of each part of the hoist 1 are described in detail below.

[0035] The motor 101 drives the hoisting machine 1 to hoist up or hoist down. The motor 101 is composed of one or more motors, and each motor is composed of, for example, a three-phase motor.

[0036] The motor control unit 132 receives as input an operation command of the hoisting machine 1 inputted through the operation input unit 202 of the operation terminal 2 , converts the operation command into a motor control command for operating the motor, and outputs the command to the motor 101 .

[0037] The motor operation information acquisition unit 131 acquires the motor operation information input to the weight mass estimation unit 133. The motor operation information is information about the operation of the motor used for weight mass estimation, such as the applied current value when the motor is driven, or the motor speed calculated based on the sensor value of the encoder, or the command value given to the motor, that is, the operating frequency, etc.

[0038] The weight mass estimation unit 133 calculates the weight mass using a weight mass estimation formula based on the motor operation information acquired by the motor operation information acquisition unit 131. In this embodiment, in the pre-shipment process of the winch, the motor operation information and weight mass data are collected, and a high-precision weight mass estimation formula is learned that takes into account individual differences in motor characteristics, etc. (hereinafter, the collected motor operation information and weight mass data set is referred to as learning data). Therefore, in the application process of the winch of this embodiment, the weight mass is only estimated using the motor operation information and the previously learned weight mass estimation formula.

[0039] In addition, the weight mass of the learning data is a value set for learning the weight mass estimation formula, and can be set manually or automatically. In addition, the value set for learning is called a reference value, which is distinguished from the value of the weight mass estimated using the weight mass estimation formula (hereinafter referred to as an estimated value).

[0040] The weight mass estimation formula uses the motor operation information as input and outputs the weight mass. The motor operation information as input can be only the current sample, or it can be time series data of a certain number of samples from the past to the present. As an example of the weight mass estimation formula, there is formula (1).

[0041] u=Wx+b…(1)

[0042] Here, x is a vector formed by connecting the motor operation information (hereinafter referred to as the motor operation information vector). u is a scalar representing the weight mass, for example, it can be an estimated value of the weight mass, or the ratio of the estimated value of the weight mass to the rated load of the product (in the latter case, if it is a product with a rated load of 1t and u=0.5, the estimated value of the weight mass is 500kg). W is a coefficient matrix representing the weights corresponding to each element of the motor operation information vector x, and b is a coefficient representing the offset.

[0043] In addition, in the weight mass estimation unit 133, a low-pass filter such as a moving average filter can be used as a post-processing to remove the time series noise of the calculated weight mass. In addition, the calculated weight mass can be output to the weight mass display unit 201 of the operation terminal 2, and further, it can also be used to prevent overload and predict life or detect grounding.

[0044] The operation terminal 2 has a heavy object mass display unit 201 and an operation input unit 202. In addition, the operation terminal 2 has hardware such as a CPU and other computing devices, a semiconductor memory and other main storage devices, a hard disk and other auxiliary storage devices and a communication device, and the description of well-known actions is appropriately omitted. In addition, the operation terminal 2 can be a teaching pendant, or it can be a smart phone or a tablet PC, a notebook PC, a desktop PC and the like. The operation terminal 2 can be provided as a dedicated terminal, or it can be provided as an application and used by the operator in their respective terminals. In addition, the operation terminal 2 exchanges information with the winch 1 through wireless communication, but it can also be wired communication. The following specifically describes the details of each part of the operation terminal 2.

[0045] The weight mass display unit 201 displays the weight mass calculated by the weight mass estimation unit 133 of the hoist 1. In addition, in the present embodiment, the weight mass display unit 201 and the operation input unit 202 are implemented in the operation terminal 2, but can also be implemented in the hoist 1, for example, the weight mass display unit 201 can also be displayed on a circuit board mounted in the hoist 1.

[0046] The operation input unit 202 is an interface for the operator to input operation instructions to the winch 1 using buttons, levers, handles or joysticks. The operation input unit 202 can be an interface that combines a display and a touch panel, or an interface implemented by software in a device such as a smart phone. The coordinates of the movement amount or the destination or the name of the place can also be input using a keyboard and a mouse. It is also possible to connect to an application management system or a production management system that manages the winches installed in the factory, and input operation instructions based on their information. Furthermore, it is also possible to have an automatic operation function that automatically inputs pre-programmed operation instructions.

[0047] For the application process of the winch of this embodiment, use Figure 3 to Figure 6 The weight of the heavy material in the winch can be estimated by applying it to the lifting time when the heavy material leaves the ground for the first time during transportation. In addition, the motor operation information and operation conditions when lifting off the ground can be used for learning. This is because only the winch is driven when lifting off the ground, and it is difficult to be affected by other external factors. Figure 3 This is a diagram to illustrate the lifting action. Figure 3 The weight 504 shown in (a) is placed on the ground and the hoist 1 is driven until Figure 3 (b) shows a state where the weight 504 is off the ground. Figure 4 1 is a diagram illustrating the positional relationship between the heavy object and the trolley when the heavy object is off the ground. When the heavy object is off the ground, it is preferred that the trolley is arranged directly above the center of gravity of the heavy object, but there are also cases such as Figure 4 The possibility of the trolley being arranged at a position deviating from the center of gravity of the heavy object. Below, the application process of the winch of this embodiment is described by taking off the ground as an example, but the same structure and process can also be used to implement it when it is on the ground.

[0048] The motor operation information acquisition unit 131 acquires the motor operation information when the vehicle leaves the ground. In order to accurately estimate the mass of the heavy object, it is required to obtain the time series data of the motor operation information. This is because if the input of the heavy object mass estimation formula can use not only the motor operation information when estimating the heavy object mass but also the past motor operation information, the expression ability of the heavy object mass estimation formula is improved. For example, the heavy object mass estimation formula that also considers the hysteresis of the motor can be obtained through learning. In the case where the data capacity is limited due to constraints such as the computing performance of the CPU or the capacity of the storage device, it is sufficient to at least obtain the data after leaving the ground.

[0049] The learning data condition may include the position deviation of the weight 504 relative to the cart 503 ( Figure 4d), or the mode of lifting during lifting. In addition, the lifting mode refers to a hoisting mode and a hoisting speed such as a high-speed or low-speed hoisting mode. The collected learning data preferably covers all conceivable off-the-ground modes in the application of the hoist 1 without omission, and as a result, the mass of the heavy object can be estimated with higher accuracy. For example, regarding the mode of the mass of the heavy object, it is preferred to collect learning data while gradually and evenly changing the mass of the heavy object from 0 kg to an overload weight (a weight exceeding the rated load of the hoist 1). In addition, the motor operation information may include back electromotive force and / or motor temperature. Thereby, the loss of the motor can be taken into account and the mass of the heavy object can be estimated more accurately.

[0050] Figure 5 This is a flowchart showing the processing flow of the hoisting machine according to this embodiment. Figure 5 The operations of the flowchart are as follows.

[0051] The processing includes steps S1 to 3 in the pre-shipment process and step S11 in the application process.

[0052] Step S1: For each machine model, under conditions where the mass of the weight, the position deviation of the weight relative to the trolley, and the lifting mode are varied, learning data (product line learning data) is collected during a lift-off test.

[0053] Step S2: Using the product line learning data collected in step S1, learn a standard heavy object mass estimation formula for each model.

[0054] Step S3: Collect learning data (each individual learning data) in the off-the-ground test during the factory test of the product, and use the newly collected data to re-learn the weight mass estimation formula for each product so that it reflects the individual differences in motor characteristics. As a result, the weight mass can be estimated with high accuracy for each product. In addition, when the individual differences between products are small, this step is not necessary.

[0055] Step S11: When the operator operates the hoist 1, the weight is estimated based on the motor operation information during operation. In the application process of this embodiment, the weight is estimated during operation using only the weight estimation formula learned in the pre-shipment process.

[0056] Figure 6 This is a flowchart showing the detailed flow of the step (S11) of estimating the mass of a heavy object during operation.

[0057] Step S301 : The operator places a heavy object on the winch 1 .

[0058] Step S302: The operator uses the operation input unit 202 to perform a lift-off operation.

[0059] Step S303 : The motor operation information acquiring unit 131 acquires the motor operation information.

[0060] Step S304: The weight mass estimating unit 133 estimates the weight mass based on the weight mass estimating formula.

[0061] Step S305 : Displaying the weight mass estimated in step S304 on the weight mass display unit 201 .

[0062] As described above, the hoist 1 estimates the weight based on the weight estimation formula in step S304, and notifies the operator of the estimated value of the weight using the weight display unit 201. The estimated weight can also be used to prevent overload, predict service life, and detect liftoff.

[0063] In this embodiment, since re-learning in the application process is not performed, a learning device is not required after shipment, and the structure of the operation terminal 2 is relatively small, so the system is simple and a low-cost winch can be provided. In addition, in this embodiment, an example of recording the ground test data during the factory test of the product in step S3 and re-learning for each product is shown, but step S3 may not be included. By performing step S3, the mass of the heavy object can be estimated with high accuracy reflecting the characteristics of each product.

[0064] [Second embodiment]

[0065] In the present embodiment, a case where a heavy object mass estimating device is installed on a hoisting machine that has already been installed or purchased will be described.

[0066] Figure 7 It is a structural diagram of the hoist and the weight mass estimation device of the present embodiment. The hoist 1 is composed of an electric motor 101. The weight mass estimation device 4 is composed of a motor operation information acquisition unit 131, a weight mass display unit 201, a learning data condition input unit 203, a weight mass estimation unit 133, a learning data setting unit 231, a learning calculation unit 331, a learning data accumulation unit 371 and a learning result accumulation unit 372. In addition, the weight mass estimation device 4 has hardware such as a CPU and other computing devices, a main storage device such as a semiconductor memory, an auxiliary storage device such as a hard disk, and a communication device, and the description of well-known actions is appropriately omitted. The following specifically describes the details of each part of the weight mass estimation device 4.

[0067] The motor operation information acquisition unit 131 acquires the motor operation information of the hoist 1. The weight estimation unit 133 calculates the weight using a weight estimation formula based on the motor operation information acquired by the motor operation information acquisition unit 131. The weight display unit 201 displays the weight calculated by the weight estimation unit 133.

[0068] The learning data condition input unit 203 is an interface for inputting the conditions of the learning data, and adopts a touch panel, a keyboard, a mouse, or the like. The learning data condition is the operating condition of the winch corresponding to the motor operating information acquired by the motor operating information acquisition unit 131, and at least includes a reference value of the weight mass. In addition, the learning data condition may include information such as the model name and the individual identification number of the winch 1. The learning data condition input by the learning data condition input unit 203 is output to the learning data setting unit 231.

[0069] The learning data setting unit 231 sets the learning data conditions input from the learning data condition input unit 203 and the motor operation information input from the motor operation information acquisition unit 131 as one set of learning data, and accumulates them in the learning data accumulation unit 371. The learning data accumulation unit 371 accumulates the learning data set by the learning data setting unit 231 of the operation terminal 2. The learning calculation unit 331 reads the learning data from the learning data accumulation unit 371 and performs learning calculation of the heavy object mass estimation formula.

[0070] The learning result accumulation unit 372 stores the load mass estimation formula outputted from the learning calculation unit 331. The load mass estimation formula is transmitted to the load mass estimation unit 133 of the hoisting machine 1 as appropriate when the load mass estimation formula is updated.

[0071] Next, the flow of the overall processing performed in the heavy object mass estimation device in this embodiment will be described. Figure 8 : is a flowchart showing the processing flow of the present embodiment. First, the weight mass estimation device 4 is installed on the winch 1 (step S20). Next, learning data consisting of motor operation information and ground-lift information is collected. Learning data collection is performed, for example, through a ground-lift test, etc. (step S1). Next, the weight mass estimation formula is learned using the learning calculation unit 331 (step S2). Then, the weight mass is estimated (step S11). Step S1, step S2, and step S11 are the same as those in the first embodiment.

[0072] In this embodiment, by installing the heavy object mass estimating device, even if there is no hoist equipped with a heavy object mass estimating function, it is possible to estimate the heavy object mass later.

[0073] [Third Embodiment]

[0074] Fig. 9 1 is a block diagram of a system related to a hoist according to the present embodiment. The hoist system is composed of a hoist 1, an operation terminal 2, and a learning device 3. The information exchange between them may be wired communication or wireless communication.

[0075] The hoist 1 has a motor 101, a motor operation information acquisition unit 131, a motor control unit 132, and a weight mass estimation unit 133, similarly to the first embodiment. However, in the hoist system of this embodiment, in the application process, not only the motor operation information is used to estimate the weight mass, but also the motor operation information and the weight mass data are collected for learning.

[0076] The motor operation information acquisition unit 131 of this embodiment outputs the motor operation information to the weight mass estimation unit 133 for estimating the weight mass when estimating the weight mass, and outputs the motor operation information to the learning data setting unit 231 of the operation terminal 2 in order to set the learning data during learning.

[0077] The weight estimation unit 133 calculates the weight using a weight estimation formula based on the motor operation information acquired by the motor operation information acquisition unit 131. Here, the weight estimation formula is transmitted from the learning result accumulation unit 372 of the learning device 3 as a learned weight estimation formula.

[0078] The operation terminal 2 of the present embodiment includes a learning data condition input unit 203 and a learning data setting unit 231 in addition to the weight display unit 201 and the operation input unit 202 .

[0079] The weight mass display unit 201 displays the weight mass calculated by the weight mass estimation unit 133 of the hoist 1. In addition, when the weight mass estimation unit 133 outputs an erroneous weight mass estimation value, a confirmation screen may be displayed to indicate whether the operator has input a correct weight mass reference value and to use the data for relearning. Relearning will be described later.

[0080] The operation input unit 202 is an interface for the operator to input operation instructions to the hoist 1. In addition, the learning data condition input unit 203 is an interface for inputting conditions of learning data. The interface of the learning data condition input unit 203 can be used in common with the interface of the operation input unit 202, or can be implemented as a different interface.

[0081] The learning data setting unit 231 associates the learning data condition input from the learning data condition input unit 203 with the motor operation information input from the motor operation information acquisition unit 131 to set a set of learning data, and transmits the set to the learning data accumulation unit 371 of the learning device 3 .

[0082] In addition, in this embodiment, the heavy object mass display unit 201, the operation input unit 202, the learning data condition input unit 203, and the learning data setting unit 231 are implemented in the operation terminal 2, but can also be implemented in the winch 1. For example, the heavy object mass display unit 201 can also be displayed on a circuit board mounted in the winch 1.

[0083] The learning device 3 includes a learning calculation unit 331, a learning data accumulation unit 371, and a learning result accumulation unit 372. In addition, the learning device 3 includes hardware such as a CPU or other computing device, a main storage device such as a semiconductor memory, an auxiliary storage device such as a hard disk, and a communication device, and the description of well-known operations is appropriately omitted. The details of each unit of the learning device 3 are specifically described below.

[0084] The learning data accumulation unit 371 accumulates the learning data set by the learning data setting unit 231 of the operation terminal 2. The accumulated data is classified into three categories. The first type is called product line learning data, which is learning data collected in advance with the learning data conditions varied for each model, and is used for learning the standard weight mass estimation formula for the model in the pre-shipment process. The second type is called individual learning data, which is learning data for each product collected during the quality assurance test at the time of the product factory test, and is used for re-learning of individual differences reflecting the characteristics of the motor in the pre-shipment process. The third type is called learning data under the application environment, which is data obtained by the operator correcting the estimated value of the weight mass calculated by the weight mass estimation unit 133 after the product leaves the factory, and is used for re-learning of the characteristics reflecting the application environment in the application process.

[0085] The learning calculation unit 331 reads the learning data from the learning data accumulation unit 371 and performs learning calculation of the weight mass estimation formula. Regarding the input learning data, the product line learning data is used in the learning of the weight mass estimation formula of the model standard in the pre-shipment process, the learning data to which the learning data of each individual is added is used in the re-learning of the individual differences reflecting the characteristics of the motor in the pre-shipment process, and the learning data to which the learning data in the application environment is added is used in the re-learning in the application process. Regarding the input learning data, as a pre-processing of the learning calculation, the time series data of the motor operation information in the data constituting the learning data can be subjected to noise addition based on Gaussian distribution, thereby increasing the variation of the data.

[0086] Next, the learning calculation process of the weight mass estimation formula performed in the learning calculation unit 331 is described. The learning calculation is, for example, a process of adjusting the coefficient matrix W and the bias amount b in the weight mass estimation formula given by equation (1) so that the weight mass u output when the motor operation information vector x is input is an estimated value with less error. Through this learning calculation, the hoist 1 can perform high-precision weight mass estimation.

[0087] First, some data are randomly extracted from the learning data input to the learning calculation unit 331. Then, for example, for the formula (1) into which the initial values ​​of the randomly determined coefficient matrix W and the initial values ​​of the bias b are substituted, the motor operation information vector x of each extracted data is input, and the error e between u calculated by the formula (1) and the reference value u' of the weight mass stored in the learning data is calculated, and the total error value E is calculated. Then, the coefficient matrix W and the bias b are slightly changed in such a way that the total error value E is smaller. The method of slightly changing the coefficient matrix W and the bias b here can use, for example, the gradient descent method. For the remaining data in the learning data input to the learning calculation unit 331, the same amount of data is randomly extracted in sequence and the same processing is repeated. When the processing of all the input learning data is completed, one cycle of the learning processing ends.

[0088] By repeating the above process, the total error value E gradually decreases, and the weight mass estimation formula is continuously updated until the learning end condition is met. The learning end condition is a condition indicating that the learning of the weight mass estimation formula is completed, such as a cycle of learning processing for the input learning data group repeated a specified number of times, or the error E reaches a specified value or less. The weight mass u output by the weight mass estimation formula obtained in this way is regarded as the optimal weight mass estimation value for the motor operation information vector x obtained at that moment.

[0089] In addition, the coefficient matrix Wa and the offset ba of the learned weight mass estimation formula are given as the initial values ​​of the coefficient matrix W and the offset b, and the learning process is called relearning. By adding new data to the learning data and learning, it is possible to update the weight mass estimation formula to be suitable for the additional learning data. In the present embodiment, by relearning using the learning data of each individual, it is possible to update the weight mass estimation formula to reflect the individual differences in the motor characteristics of the product. In addition, by relearning using the learning data in the application environment, it is possible to update the weight mass estimation formula to reflect the characteristics of the application environment. In this way, a more accurate weight mass estimation that reflects the characteristics of each product and the characteristics of the application environment in which the product is used can be performed.

[0090] In addition, the weight mass estimation formula including the learned coefficient matrix W and the offset b is stored in the learning result accumulation unit 372.

[0091] The learning result accumulation unit 372 stores the weight mass estimation formula output by the learning calculation unit 331, and appropriately transmits the weight mass estimation formula to the weight mass estimation unit 133 of the hoist 1. In addition, the standard weight mass estimation formula for each model of the product, or the weight mass estimation formula for each individual or different application environments can also be accumulated in advance in the learning result accumulation unit 372. Then, these weight mass estimation formulas can be appropriately referenced by specifying the model name of the product or the individual identification environment or application environment.

[0092] This allows the operator of the hoisting machine 1 to easily select and use the weight mass estimation formula in accordance with the application environment.

[0093] Fig.10 This is a flowchart showing the processing flow of the hoisting machine system according to this embodiment. Fig.10 The operations of the flowchart are as follows.

[0094] The process is composed of steps S1 to 3 in the pre-shipment process and steps S11 to 15 in the application process. Steps S1 to S3 in the pre-shipment process are the same as those in the first embodiment, so only steps S11 to 15 in the application process will be described.

[0095] Step S11: When the operator operates the hoist 1, the weight of the heavy object is estimated based on the motor operation information during the operation.

[0096] Step S12: If the operator determines that the estimated value of the weight mass calculated in step S11 is wrong, the operator can correct the estimated value. In this step, the operator is asked to determine whether correction is required via the weight mass display unit 201. If the operator determines that correction is required, the process moves to step S13. If correction is not required, the process moves to step S14.

[0097] Step S13: The learning data setting unit 231 sets the reference value of the weight corrected by the operator and the motor operation information and operation conditions at the time of leaving the ground as relearning data (learning data under application conditions). Then, the relearning data is transmitted to the learning data accumulation unit 371.

[0098] Step S14: Check whether to perform re-learning. The judgment criterion may be an instruction from the operator or a situation where a certain amount of re-learning data has been accumulated. If "yes", the process moves to step S15. If "no", the process ends.

[0099] Step S15: Relearning is performed using the relearning data collected in step S13 so that the relearning reflects the characteristics of the application environment. As a result, the characteristics of the application environment can be reflected and the mass of the heavy object can be estimated with high accuracy. In addition, this step is not necessary when the change in the motor characteristics caused by the characteristics of the application environment is small.

[0100] This embodiment uses the learning data collected in steps S2, 3, and 15 to learn the weight mass estimation formula, and can update the weight mass estimation formula according to the tendency of the learning data. Therefore, the characteristics of the product and the application environment can also be reflected in the weight mass estimation formula, and as a result, a more accurate weight mass can be estimated. In addition, for winches of different models, even if the corresponding models are not constructed, the weight mass can be estimated by only collecting learning data through the same process.

[0101] Fig.11 This is a flowchart showing a detailed flow of the steps (S1 and S3) of collecting data on motor operation information and learning data conditions through a ground-lift test in a pre-shipment process.

[0102] Step S101: The manufacturer determines the conditions for collecting learning data in advance. The conditions for collecting learning data are the conditions of the combination of the weight mass in the lift-off test, the position deviation of the weight relative to the trolley, and the lifting mode, and are preferably set in a manner that covers all conditions that can be imagined on the application winch 1. Then, the manufacturer sets the weight on the winch 1 according to the conditions for collecting learning data.

[0103] Step S102: The manufacturer inputs a lift-off operation using the operation input unit 202, and records the time series data of the motor operation information during lift-off. The time series data of the motor operation information is preferably from the start of lift-off until the weight becomes stable.

[0104] Step S103: The manufacturer uses the learning data condition input unit 203 to input the learning data conditions.

[0105] Step S104 : The learning data setting unit 231 sets the time series data of the motor operation information recorded in step S102 as learning data in association with the learning data condition set in step S103 .

[0106] Step S105 : The learning data set in step S104 is transferred to the learning data accumulation section 371 .

[0107] As described above, in this embodiment, the weight of the heavy object can be accurately estimated by learning only by collecting data from multiple lift-off tests. In addition, this operation has the advantage of being independent of humans and can be performed even without advanced knowledge of products such as hoist motors.

[0108] Fig.12 This is a flowchart showing the detailed flow of the steps (S2, S3 and S15) of learning the heavy object mass estimation formula.

[0109] Step S201 : Reading learning data from the learning data accumulation unit 371 .

[0110] Step S202: Use the learning calculation unit 331 to learn the heavy object mass estimation formula.

[0111] Step S203 : The weight mass estimation formula outputted in step S202 is accumulated in the learning result accumulation unit 372 .

[0112] Step S204 : The weight mass estimation formula is transmitted from the learning result accumulation unit 372 to the weight mass estimation unit 133 .

[0113] As described above, the learning process of the heavy object mass estimation formula is performed in a manner reflecting the learning data collected in steps S1 , 3 , and 13 .

[0114] Fig.13 This is a diagram showing an example of a system in which a plurality of hoisting machines 1 exist.

[0115] like Fig.13 As shown in the example, when the system has multiple hoists 1A to 1C, the same operation terminal 2 can be used to operate and set learning data. Then, the learning data collected in the hoists 1A to 1C are aggregated and accumulated in the same learning device 3. Then, the learning device 3 performs learning based on these data. As described above, multiple hoists 1A to 1C can be processed using the same operation terminal 2 and learning device 3, which can reduce the burden on the operator.

[0116] Fig.14 This is an example of a system in which the learning device exists on a cloud server.

[0117] like Fig.14 As shown, the learning device 3 can be set on a cloud server. In addition, the cloud server can be a computing device used by the winch manufacturer. The cloud server may not be Fig.14 Instead of being connected to one hoist 1 and an operating terminal 2 as shown, it is connected to a plurality of hoists 1 and operating terminals 2. Furthermore, it is also possible to Fig.13 As shown, a plurality of hoists 1 are further connected to the operator terminal 2 .

[0118] If as described above, the operator does not need to set up, apply, and maintain the learning device separately. In addition, if the data from multiple operators are aggregated to a cloud server, learning can be performed using a large amount of data, and a more accurate weight mass estimation formula can be provided.

[0119] As described above, according to this embodiment, it is possible to provide a method for estimating the mass of heavy objects with high accuracy and a winch equipped with the method, which does not require additional sensors and which collects and learns the data of the motor operation information and reference value of the mass of heavy objects when off the ground for each model, product, and application environment, and which can also take into account the individual differences in the motor characteristics and the characteristics of the application environment.

[0120] The first to third embodiments described above are described in detail for the purpose of explaining the present invention in an easy-to-understand manner, and are not limited to all structures that must be described. In addition, a part of the structure of a certain embodiment can be replaced with the structure of another embodiment, and the structure of another embodiment can be added to the structure of a certain embodiment. In addition, for a part of the structure of each embodiment, other structures can be added, deleted, or replaced.

[0121] Description of Reference Numerals

[0122] 1. 1A~1C winch

[0123] 2 Operation terminal

[0124] 3 Learning Devices

[0125] 4 Heavy object mass estimation device

[0126] 101 Electric Motor

[0127] 131 Motor operation information acquisition unit

[0128] 132 Motor control unit

[0129] 133 Heavy object mass estimation department

[0130] 201 Heavy object mass display unit

[0131] 202 Operation input unit

[0132] 203 Learning data condition input unit

[0133] 231 Learning data setting unit

[0134] 331 Learning Computing Department

[0135] 371 Learning Data Accumulation Department

[0136] 372 Learning results accumulation section.

Claims

1. A winch system, comprising a winch capable of moving a suspended weight in a vertical direction, an operation terminal for operating the winch, and a learning device for learning a weight mass estimation formula used when estimating the weight of the weight, wherein the winch system is characterized in that: The hoist includes an operation information acquisition unit for acquiring operation information of the hoist, and a weight mass estimation unit for estimating the weight of the hoist from the operation information acquired by the operation information acquisition unit. The operation terminal includes a weight display unit that displays the weight of the weight calculated by the weight estimation unit of the hoist. The learning device comprises a learning data accumulation unit for accumulating learning data required for learning the heavy object mass estimation formula, a learning calculation unit for reading the learning data from the learning data accumulation unit to perform learning calculations, and a learning result accumulation unit for accumulating the heavy object mass estimation formula output by the learning calculation unit. When the operator determines that the weight displayed by the weight mass display unit is wrong, the weight mass corrected by the operation terminal is associated with the operation information and accumulated in the learning data accumulation unit as re-learning data.

Images