Compressor control system and compressor control method

Abstract

According to the present invention, the pressure response to each flow rate at a plurality of terminals is measured to construct a machine learning model for predicting fluctuations in terminal pressure, and thereby optimal air compressor control is realized. In this compressor control method for a pneumatic system having a plurality of terminals that consume compressed air, a control unit of a system 400: (a) measures supply side pressures by means of pressure gauges 305 to 307 on a discharge side from which compressed air is supplied; (b) uses pressure gauges 113, 123, ..., 243, flowmeters 112, 122, ..., 242, and valves 111, 121, ..., 241 provided in some of the terminal ends of pipes 11, 12, ..., 24, to acquire pressure values and flow rate values when the valves are opened or closed; and (c) uses supply pressure values from air compressors 2 to 4 and demand side pressure values and flow rates to train a terminal pressure prediction model for predicting the pressure at each of the terminals 11, 12, ..., 24. During actual operation of the pneumatic system, the control unit inputs the supply pressure values and terminal air flow rate values into the terminal pressure prediction model to calculate the optimum discharge pressures of the air compressors.

Description

Compressor control system and compressor control method

[0001] The present invention relates to a compressor control system and a control method for a facility that has an air compressor controlled by a variable speed device and supplies compressed air to compressed air-using equipment connected to the end via multiple pipes, and in particular to a system that efficiently controls the compressor taking into account fluctuations in the amount of compressed air used in the facility.

[0002] Pneumatic equipment that supplies compressed air to various parts of a factory or other facility temporarily stores compressed air compressed by one or more air compressors in a receiver tank (air tank), and then supplies it from this receiver tank via a piping system and pneumatic equipment (filters, dryers, control valves, etc.) to equipment that uses compressed air in the factory's production process, such as air cylinders and air blowers. The pressure loss in the air piping from the air compressor's discharge port to the terminal increases the closer to the terminal, but pressure loss also changes depending on fluctuations in the air discharge volume of the air compressor and the amount of air used at the terminal. While the number of terminals in a piping network may be as few as a few, large factories may have dozens or even hundreds of terminals, and pressure loss constantly changes depending on fluctuations in the amount of air discharged at each terminal.

[0003] In order to maintain the pressure of compressed air supplied to multiple terminals at a predetermined level or higher, conventional pneumatic equipment has set the discharge pressure of air compressors higher, anticipating maximum pressure loss. However, the amount of air used at the terminals often fluctuates depending on the type of compressed air-using equipment and its operating rate. Furthermore, the amount of air used and the associated pressure fluctuations often vary from day to day. Normally, when the amount of air used is small, the pressure loss decreases, resulting in unnecessary high pressure at the terminals, resulting in unnecessary waste in the operation of the air compressor. In particular, if the discharge pressure of the air compressor is kept constant, the discharge pressure will be higher than necessary when the amount of air used is low, resulting in excessive power consumption.

[0004] Conventional techniques have been used to stabilize the discharge pressure of compressed air while achieving energy savings by controlling the discharge pressure of the air compressor in response to fluctuations in the amount of air used at the end of the line. This involves adjusting the discharge pressure of the air compressor so that the supply pressure to the end of the line is constant. To achieve this control, it is necessary to monitor the pressure at the end of the line and understand the pressure transmission delay from the compressor's discharge port to the end of the line. A typical approach involves installing pressure gauges at the end of the line and building a physical model of the air piping (a simulator using differential fluid calculations) to calculate the end pressure. However, in large-scale facilities where the end pressures to be predicted span dozens to hundreds of locations, the number of installed pressure gauges becomes enormous, and the piping diagrams required for building the physical model are often not available, resulting in significant labor costs for checking and measuring the piping. This can make this approach unsuitable for actual control of large-scale pneumatic facilities in factories and other facilities during actual operation.

[0005] One approach to reducing the number of pressure gauges to be installed and understanding the terminal pressure and transmission delay without using a piping diagram is to model the characteristics of air piping from data.As a related technique, Patent Document 1 proposes a method for detecting faults in a piping network using a sensor that determines the physical parameters of the gas.

[0006] Special Publication No. 2022-517504

[0007] In the technology of Patent Document 1, adjustable throttle valves in a piping network are opened and closed in a predetermined sequence, and measurements are taken with two groups of pressure gauges (and flow meters, if necessary) during the opening and closing. Based on the difference in measurements between the first and second groups, a mathematical model is constructed to identify and quantify partial or total blockages in each piping network or increases in resistance within each pipe. While this model represents the characteristics of an air piping network, it does not have the function to directly calculate the end pressure or the transmission delay from the discharge port to the end, making it difficult to use for controlling a compressor based on the end pressure.

[0008] An object of the present invention is to realize a compressor control system and method that provides pressure gauges and flow meters at multiple terminals and measures the pressure response to each flow rate to thereby construct a machine learning model in advance that predicts the end pressure from the flow rate. Another object of the present invention is to realize a compressor control system and method that, during actual operation of pneumatic equipment, zones the piping network, predicts the end pressure by inputting air flow rate values ​​at a representative terminal or all terminals in each zone into a machine learning model, and optimally controls the discharge pressure of the air compressor based on the predicted value. Yet another object of the present invention is to realize a compressor control system that enables a user who sets the number of operating air compressors and their discharge pressures to simulate the discharge pressure of the air compressor and the end pressure of the zoned zones using the constructed machine learning model.

[0009] Representative features of the invention disclosed in this application are as follows. According to one feature of the present invention, a compressor control system controls equipment that supplies compressed air discharged from an air compressor via multiple pipes to multiple terminals that consume the compressed air. The system includes a pressure gauge at one or multiple supply points near the air compressor or a receiver tank, a pressure gauge, a flow meter, and a valve connected to all or some of the terminals of the pipes, and stores data of the actual pressure and flow rate measured during the operation in a storage device. Using the measured data, a terminal pressure prediction model construction unit learns a terminal pressure prediction model that inputs the pressure at the supply point and the flow rates at the multiple terminals and outputs the terminal pressure. During actual operation of the equipment, the terminal pressure prediction unit predicts the terminal pressure by inputting the measured pressure values ​​at the supply points and flow rate information for each terminal into the trained terminal pressure prediction model during actual operation of the equipment. The set pressure search unit then searches for and sets the optimal discharge pressure of the air compressor using the predicted terminal pressure values. One or more air compressors are provided, and the operation control unit controls the rotation speed and number of operating air compressors so that the multiple terminal pressures set by the set pressure search unit fall within a predetermined range.

[0010] According to another feature of the present invention, a terminal pressure prediction unit is used to control the rotation speed and number of operating air compressors so that the future predicted values ​​of multiple terminal pressures and the air compressor discharge pressure remain constant. A terminal pressure prediction model is used to calculate this predicted value, and the terminal pressure prediction model construction unit uses many terminal pressure measurements and flow rate information during model learning to set the optimal discharge pressure. During actual operation, the terminal pressure is predicted using only a smaller number of terminal flow rate information than during learning. Furthermore, multiple terminal sections branching off from multiple pipes (pipe groups) are defined as zones, and a representative terminal for measuring flow rate is set for each zone. The terminal pressure prediction unit sets the optimal discharge pressure of the air compressor using only the actual measurement data of representative terminal flow rates. As a result, during equipment operation, the air compressor can be controlled by using the terminal pressure prediction model with flow rate measurement data from a small number of terminals as input.

[0011] According to yet another feature of the present invention, an input / output unit is provided for inputting and outputting (displaying) information, and a user inputs the rated flow rate, operation method, and availability rate of the equipment connected to the terminal, which are then stored in advance in a storage device as a model database. The terminal pressure prediction unit estimates a change pattern in the terminal flow rate from the change pattern in the input information and the measured value of the supply point pressure, and predicts the measured value of the supply point pressure and a future pressure change pattern. A display device is also provided for visually displaying information to the user, and the optimal discharge pressure of the air compressor set by the set pressure search unit is displayed on the display device, allowing the user to monitor the operating status of the equipment. In addition, in the compressor control system of the present invention, the set pressure search unit may simulate the optimal discharge pressure of the air compressor in response to the user's input of the terminal target pressure, and display the simulated optimal discharge pressure on the display device.

[0012] According to yet another feature of the present invention, there is provided a compressor control method for a pneumatic system in which compressed air discharged from an air compressor is supplied to a plurality of terminals consuming the compressed air via a plurality of pipes, using a control unit having a processor, and an end pressure prediction control system having an input / output device and a memory device connected to the control unit, wherein the control unit: a) measures the supply side pressure using one or more pressure gauges provided on the discharge side that supplies compressed air to the plurality of pipes; b) obtains pressure values ​​and flow rate values ​​when the valves are open or closed using pressure gauges, flow meters, and valves provided on all or some of the ends of the pipes; c) learns an end pressure prediction model having a pressure for each end as an output using the supply pressure value from the air compressor and the pressure value and flow rate value on the demand side; and when the pneumatic system is in operation, the control unit: d) connects compressed air utilizing equipment to the end instead of the valve; and e) calculates an optimal discharge pressure of the air compressor by inputting the supply pressure value and the end flow rate value into the end pressure prediction model.

[0013] According to the present invention, it is possible to monitor and predict the pressure margins at multiple terminals according to the usage status of compressed air, and by controlling the air compressor so as to maintain the pressure at the terminal with the smallest pressure margin above the lower limit, it is possible to reduce the discharge pressure on the air compressor side as much as possible. As a result, it is possible to realize a compressor control system that contributes to energy savings and is environmentally friendly. In addition, it is possible to effectively suppress hunting caused by a delay in the fluctuation of the air compressor discharge pressure to the terminal pressure. Issues, configurations, and effects other than those described above will become clear from the description of the following embodiments.

[0014] 10 is a schematic configuration diagram of pneumatic equipment to which an embodiment of the present invention is applied, showing an example of a piping network for sending out compressed air from an air compressor 2. FIG. 11 is a configuration diagram of a compressor control system 1 according to an embodiment of the present invention. FIG. 12 is a flowchart showing the procedure for measuring the pressure and flow rate at each terminal of FIG. 2. FIG. 13 is a flowchart showing the procedure for building a model for terminal pressure predictive control according to this embodiment. FIG. 14 is an example of a display screen when building a model in steps 511 to 512 of FIG. 4. FIG. 15 is a system configuration diagram when actual equipment is operated using a terminal pressure predictive model when all terminals are measured in this embodiment. FIG. 16 is a diagram showing an example of an input / output screen 470 in the input / output unit 410 of FIG. 6. FIG. 17 is a diagram showing an example of a method for estimating an optimal pressure taking multiple terminals into consideration in the terminal pressure predictive control system 400 of FIG. 2. FIG. 18 is a diagram showing an example of control for suppressing hunting due to a delay in terminal pressure fluctuations in the terminal pressure predictive control system 400 of FIG. 2. FIG. 19 is a flowchart of compressor control in the terminal pressure predictive control system 400 of FIG. 2. FIG. 19 is a subroutine showing the procedure for searching for the optimal value of the discharge pressure in step 523 of FIG. 10. FIG. 19 is a configuration diagram showing the model operation configuration during zoning in the terminal pressure predictive control system 400 according to a second embodiment of the present invention. 12. It is an example of a definition table when the terminals of the control object are assigned to each zone in FIG. 12. It is a schematic configuration diagram of pneumatic equipment to be modeled at the time of zoning in the terminal pressure prediction control system 400 according to a modified example of the second embodiment of the present invention. It is a diagram showing a model for predicting terminal pressure including unmeasured terminals using learned model parameters. It is a list showing measurement points and operating characteristics in three usage forms of the compressor control system according to the present invention. It is a configuration diagram at the time of compressor setting recommendation in the compressor setting pressure recommendation system 400A according to the third embodiment of the present invention. It is a display screen at the time of compressor setting recommendation in the compressor setting pressure recommendation system 400A according to the third embodiment of the present invention.

[0015] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the following drawings, the same parts are given the same reference numerals and repeated explanations will be omitted.

[0016] 1 is a schematic diagram of a pneumatic facility to which an embodiment of the present invention is applied, showing part of a piping network for sending out compressed air from an air compressor 2. In this figure, compressed air is supplied from the supply-side facility to the left of the dashed line via a plurality of main pipes 300, 10, 20 to terminals 11-14, 21-24. One or more air compressors 2 are provided in the supply-side facility, and compressed air is sent through the supply-side main pipe 300 to a branch point 310. A pressure gauge 305 is provided in the main pipe 300, and the discharge pressure P ｃｏｍｐ is measured. The piping structure downstream of branch point 310 (the side farthest from the air compressor in terms of the air flow) varies depending on the factory and type of pneumatic equipment. Furthermore, because the amount of compressed air used in pneumatic equipment varies, the number of air compressors used in the supply-side equipment may be more than one. When multiple air compressors are used, a receiver tank (not shown) may be provided upstream of branch point 310 (the supply-side equipment closer to air compressor 2).

[0017] Downstream of branch point 310, a main pipe 10, for example, having a straight pipe shape with a dead end, is provided. Multiple thin branch pipes (not shown) branch off from the main pipe 10, and the ends of the branch pipes are termed ends 11 to 14. Hereinafter, the end portions of the branch pipes are referred to as "ends." Although not shown, devices using high-pressure air are provided downstream of the ends 11 to 14. Valves 111, 121, 131, and 141 are provided downstream of the ends 11 to 14 to open or close the air flow. Valves 111, 121, 131, and 141 are not necessarily provided; the ends 11 to 14 may be directly connected to the high-pressure air-using devices without the valves 111, 121, 131, and 141, or they may be configured to be connectable and detachable using couplers. Here, the user defines the main pipe 10 and the portion of the ends 11 to 14 as "Zone 1." The portion of piping 11 surrounded by a dashed line has been selected by the user as a representative piping that effectively reflects pressure fluctuations in zone 1, and is managed as "end #11."

[0018] The header 20 is a pipeline that branches off from a branch point 310 and is then looped. Terminals 21 to 24 are provided, branching off from the header 20. Here, the user defines the header 20 and the terminals 21 to 24 as Zone 2. Note that the headers 10 and 20 can have any number of terminals, and the number of terminals within the headers 10 and 20 is not limited to four as shown. Downstream of terminals 21 to 24 are provided equipment (not shown) that uses high-pressure air, and valves 211, 221, 231, and 241 that open or close the air flow. The portion of the piping 21 surrounded by a dashed line is selected by the user as a representative piping that reflects pressure fluctuations in Zone 2 and is managed by this system as "Terminal #21." A flow meter 212 and a pressure gauge 213 are provided within Terminal 21.

[0019] The main pipes 10 and 20 from branch 310 of main pipe 300 and the terminals connected to them are considered as one "zone", and a hypothesis is made that the pressure fluctuation characteristics (time response of pressure to changes in the amount of air used) are roughly the same within the same zone. One terminal (11, 21) is defined as a representative point for each zone, and this terminal is always used as the measurement point when creating the machine learning model. When creating the machine learning model, pressure gauges 113 and 213 and flow meters 112 and 212 are installed at the representative points, and the flow rate and pressure are measured while changing the opening of valves 111 and 211. By this operation, the terminal pressure p is associated with the terminal number, and the pressure p of terminals 11 and 21 is calculated. １ , p ２ , terminal flow rate u １ , u ２are measured and recorded in the system's storage device. In this way, the basic idea of ​​the first and second embodiments is to learn the pressure fluctuation characteristics of each zone from the measurement data of the representative terminals 11 and 21, and to monitor and control the pressure at each terminal, including its future value, using the pressure fluctuation characteristics and flow rate information for each of the multiple terminals including the representative terminals 11 and 21. The number of representative terminals is arbitrary, and one or more terminals may be set for each zone (Example 2). Alternatively, a machine learning model for terminal pressure predictive control may be constructed by measuring the pressures and flow rates of all or almost all terminals without setting a representative terminal (Example 1). Once the machine learning model for terminal pressure predictive control (terminal pressure predictive model) is constructed, optimal operation control of the air compressor is performed using the constructed terminal pressure predictive model when the actual pneumatic system is in operation.

[0020] 2 is a configuration diagram of a compressor control system 1 according to a first embodiment of the present invention. In the compressor control system 1, a pneumatic facility having one or more air compressors 2 to 4 and a piping network 8 extending from the air compressors 2 to 4 to a supply destination device (not shown) is controlled, and a plurality of measurement data (piping pressure, piping flow rate) is acquired from the controlled object, collected by a terminal pressure prediction control system 400, and processed.

[0021] The left side of Figure 2 illustrates the pneumatic equipment shown in Figure 1 arranged so that the flow of air from the supply source (air compressors 2, 3, ..., 4) to supply equipment (not shown) connected to valves 111, 121, ..., 241 flows from top to bottom. Piping network 8 includes supply-side header 300 shown in Figure 1 and multiple headers 10, 20, etc. connected to supply-side header 300, and also includes a receiver tank (not shown) for mitigating fluctuations in discharge pressure to the multiple headers. The supply-side equipment includes multiple air compressors 2, 3, ..., 4. The number and type of air compressors are arbitrary, as long as they can stably supply the required amount of compressed air to the supply destination (high-pressure air-using equipment) (not shown).

[0022] A plurality of pipes 11, 12, ..., 24 branch off from the piping network 8. These pipes 11, 12, ..., 24 correspond to the terminals 11, 12, ..., 24 shown in Fig. 1, respectively. For ease of explanation, Fig. 2 illustrates the piping network 8 and the terminals 11, 12, ..., 24 separately, but in a broad sense, the terminals 11, 12, ..., 24 are also part of the piping network 8. Individual pipes 302, 303, ..., 304 connect the air compressors 2, 3, ..., 4 to the piping network 8. Pressure gauges 305, 306, ..., 307 are provided on the pipes 301, 302, ..., 304, respectively, and their outputs are transmitted to the terminal pressure prediction and control system 400 in real time. As explained in the legend at the bottom left of the figure, the dashed arrows extending from the control devices and meters indicate "data flow" via signal lines or wireless communication, and the solid arrows extending from users 80 and 380 indicate input instructions by "user operations." The equipment on the left side of Figure 2 controls a wide variety of devices other than the air compressors 2 to 4 and is formed as an "air pressure system."

[0023] In this embodiment, the actual characteristics of the pneumatic system are measured in advance to construct (machine learning) an end pressure prediction model to be used during actual operation. In an actual pneumatic system such as a factory or facility, pressure gauges (113, 123, ..., 243), flow meters (112, 122, ..., 242), and valves (111, 121, 241) are installed at all ends (ends of pipes 11, 12, ..., 24). Note that installing flow meters at all ends does not strictly mean installing measuring devices at 100% of ends; measurements at the ends of small pipes that do not affect pressure fluctuations or the ends of pipes that are physically difficult to measure may be omitted. These tasks are performed in advance by a user 380 before constructing the end pressure prediction model. The user 380 is, for example, a serviceman from the manufacturer that provides the end pressure prediction control system 400. The portions where the valves (111, 121, 241) are provided are usually connected to high-pressure air-using equipment (not shown) to which compressed air is supplied, and predetermined operations, tasks, controls, etc. are performed using the compressed air. However, when making measurements to create a model, the terminals are disconnected from the high-pressure air-using equipment, and the downstream side of the valves (111, 121, 241) is opened to the atmosphere or closed. Then, a user 380 measures changes in pressure and flow rate while opening and closing each valve at each terminal, and records the measured data in an experiment DB (database) 420 of the terminal pressure prediction and control system 400.

[0024] The terminal pressure prediction control system 400 has a processing device (processor) such as a CPU (Central Processing Unit), DSP (Digital Signal Processor), GPU (Graphics Processing Unit), FPGA (Field-Programmable Gate Array), etc., a storage device (memory) such as a ROM (Read Only Memory), RAM (Random Access Memory), HDD (Hard Disk Drive), SSD (Solid State Drive), etc., an input / output unit that inputs information from the user and displays information to the user, a network interface, etc., and is configured so that a computer program stored in advance in the storage device can be executed by the processor. This device can be realized by, for example, a personal computer, a portable terminal device, or a server device. The terminal pressure prediction control system 400 may also be configured to be incorporated into a control panel (not shown) for controlling the number of air compressors.

[0025] A processor (not shown) of the terminal pressure prediction control system 400 executes a specific computer program to execute the machine learning function of the terminal pressure prediction model construction unit 401 to generate a terminal pressure prediction model 430, and then executes the prediction function of the terminal pressure prediction unit 402 based on the model. The input / output unit 410 has an output function for visually displaying various information to the user 80 and an input function for accepting input instructions from the user 80. The input / output unit 410 can be configured with a dot-matrix display device and input devices such as a keyboard and a mouse, but it can also be configured as both an input device and an output device using a touch-type liquid crystal display panel. The detailed configuration of the input / output unit 410 is optional, and a remote operation terminal, mobile terminal, smartphone, etc. can also be used as an input / output device. The experiment DB (database) 420 is constructed in a non-volatile storage device such as an HDD or SSD, and records measurement data acquired from each pressure gauge and each flow meter when the user 380 operates the valves 111, 121, ..., 241, along with the acquisition time. The model DB (database) 430 stores data of the terminal pressure prediction model constructed as a result of learning by the terminal pressure prediction model construction unit 401, and is stored in a non-volatile storage device such as an HDD or SSD.

[0026] The terminal pressure prediction model construction unit 401 constructs a machine learning model that predicts fluctuations in pressure at the terminals 11, 21, ..., 24, using the discharge pressures measured by the pressure gauges 305 to 307 and the pressure and flow rate measurements measured at each terminal 11, 21, ..., 24. The constructed machine learning model is stored in a model database (DB) 430. The terminal pressure prediction unit 402 inputs the flow rate values ​​and discharge pressure measurements measured at each terminal into the terminal pressure prediction model during operation of the pneumatic system, and predicts the pressure at the terminals 11, 21, ..., 24. The set pressure search unit 403 searches for set pressures for the air compressors 2 to 4 that approach the target pressures at each terminal set by the user 80, based on the predicted values ​​predicted by the terminal pressure prediction unit 402. The set pressure set by the set pressure searching unit 403 is transmitted to the number of operating units / rotation speed operating unit 404, which controls the number of operating units and rotation speed of the air compressors 2 to 4 so as to realize the set pressure. The terminal pressure predicting unit 402, the set pressure searching unit 403, and the number of operating units / rotation speed operating unit 404 are realized by a processor (not shown) executing a computer program for realizing the above functions.

[0027] Figure 3 is a flowchart showing the procedure for measuring the pressure and flow rate at each end of the piping network targeted for end-point pressure predictive control shown in Figure 2. First, user 380 disconnects the high-pressure air-using equipment from the ends of piping 11-14 and 21-24 and connects valves 111, 121, ..., 241, flow meters 112, 122, ..., 242, and pressure gauges 113, 123, ..., 243 (step 501). The downstream side of the valves (111, 121, ..., 241) is open to the atmosphere. User 380 performing this operation could be, for example, a service technician from the manufacturer of air compressors 2-4 or a control engineer for end-point pressure predictive control system 400. However, the distinction between users 80 and 380 is not strict, and the same person could perform the operations. Next, user 380 selects the end of piping 11-14 and 21-24 to be operated (step 502).

[0028] Next, the user 380 sets the opening of the selected terminal valve to 100% and waits for about one minute (step 503). During this time, the outputs of the flow meters 112, 122, ..., 242, the pressure gauges 113, 123, ..., 243, and the pressure gauges 305, 306, and 307 are electrically transmitted to the terminal pressure predictive control system 400 via signal lines. A processor (not shown) of the terminal pressure predictive control system 400 stores the transmitted outputs in the experiment DB 420. Next, the user 380 sets the valve opening to 50% and waits for about one minute (step 504). During this time, the pressures measured by the flow meters 112, 122, ..., 242, the pressure gauges 113, 123, ..., 243, and the outputs of the pressure gauges 305, 306, and 307 are electrically transmitted to the terminal pressure predictive control system 400 via signal lines and stored in the experiment DB 420. It is also possible to configure the system so that the measured data is temporarily stored in a storage device built into the pressure gauges 113, 123, ..., 243 and the flow meters 112, 122, ..., 242, and after the measurement is completed, the stored data is copied to a USB memory (USB is a registered trademark of USB Implementers Forum), and then the removed USB memory is connected to the terminal pressure prediction and control system 400 and transferred to the experiment DB (database) 420. Furthermore, it is also possible to configure the system so that the outputs of the pressure gauges 113, 123, ..., 243 and the flow meters 112, 122, ..., 242 are stored in a measurement personal computer via a signal line, and after the measurement is completed, the data is transferred to the experiment DB (database) 420 via wireless communication, a connection via USB (registered trademark), or the like to the terminal pressure prediction and control system 400.

[0029] Next, the user 380 determines whether there are any unoperated terminals, i.e., whether the outputs from the flowmeter and pressure meter at valve openings of 100% and 50% have been stored in the experiment database 420 of the terminal pressure prediction and control system 400 (step 505), and if there are any unoperated terminals, the process returns to step 502 and repeats steps 502 to 504. The determination in step 505 may not only be made by the user 380, but also by a processor (not shown) of the terminal pressure prediction and control system 400 executing a computer program to display a list of terminals to be operated on the screen of the input / output unit 410, sequentially specify the operation of the valves of the terminals to be measured on the screen, and when the measurement of the specified terminal is completed, display the terminal as completed, thereby assisting the user 380 in opening and closing the valves.

[0030] In step 505, when the measurement of pressure data and flow rate data for all the terminals (pipes 11 to 14, 21 to 24) to be operated is completed, the data is compiled to construct an experimental database 420, and stored in the storage device of the terminal pressure predictive control system 400 (step 506). As described above, in this embodiment, basically, valves, flow meters, and pressure gauges are provided at the ends of all of the pipes 11 to 14, 21 to 24 to measure actual values, but installation may be omitted for small pipes that have little effect on the terminal pressure predictive control system, or for pipes where installation of valves, flow meters, or pressure gauges is difficult, so it is not necessarily necessary to cover all pipes.

[0031] When data measurement for predictive control according to this embodiment is completed and storage of measurement values ​​in the experiment database 420 is completed according to the procedure shown in the flowchart of Fig. 3, a processor (not shown) of the terminal pressure predictive control system 400 constructs a terminal pressure predictive model according to the procedure shown in the flowchart of Fig. 4. The procedure shown in Fig. 4 is executed by a terminal pressure predictive model construction unit 401. The terminal pressure predictive model construction unit 401 is realized by software when the processor of the terminal pressure predictive control system 400 executes a learning program for constructing a terminal pressure predictive model that is stored in advance in a storage device (not shown).

[0032] First, the user 80 operates the input / output unit 410 (e.g., a keyboard or mouse) to specify parameters and a data range for model construction (step 510). Next, the terminal pressure prediction model construction unit 401 constructs a terminal pressure prediction model from the pressure data and flow rate data within the specified data range (step 512). Specifically, the terminal pressure prediction model has a function for inputting the actual measured values ​​of the terminal flow rate and discharge pressure into the terminal pressure prediction model to predict the terminal pressure. As a basic configuration of the terminal pressure prediction model, a model is considered in which the past history of discharge pressure and the square of the terminal flow rate are input and the terminal pressure is predicted up to a time point L (≧1) in the future, taking into consideration that the total amount of piping used can be estimated from the time rate of change of the discharge pressure and that pressure loss due to friction within the pipe is proportional to the square of the flow velocity (Darcy-Weisbach equation). This terminal pressure prediction model is shown in Equation 1.

[0033] Here, P ｃｏｍｐ (t), ..., P ｃｏｍｐ (t-d) is the discharge pressure of the air compressor at time t, ..., time t-d. U １ ２ (t), ..., U Ｍ ２ (t) is the flow rate at the terminals #1 to #M at time t, and Δt is the sampling rate (assumed to be about 1 second). １，１ ... ] is the matrix coefficient A, and [b １，１ ... ] is the matrix coefficient B.

[0034] The role of each matrix in Equation 2 is to express how much delay there is in the transmission of the change in discharge pressure over time to the end pressure.

[0035] Furthermore, Equation 3 expresses the degree of pressure loss caused by the terminal flow rate and the delay with which it is transmitted.

[0036]

[0037] The model construction algorithm is as follows: (1) Select the number of prediction points L and the number of delay points d. The number of prediction points L is selected so that LΔt is in the same order as the delay (approximately 10 to 20 seconds if Δt = 1 second). The number of delay points d is selected so that d>M(L-1) in order to successfully solve the optimization problem in the subsequent stage. (2) At time t１ , ..., t Ｎ From the data of the supply pressure, end pressure, and end flow rate around the time t, the matrix Y of the following equation 4, the matrix X of the following equation 5, and the matrix U of the following equation 6 are constructed. １ , ..., t Ｎ However, each time point t ｉ For the interval [t ｉ -dΔt,t ｉ +LΔt] must be measured continuously.

[0038]

[0039]

[0040] (3) In order to obtain the matrix R shown in Equation 7, the optimization problem shown in Equation 8 is solved.

[0041]

[0042] Let λ be an extremely small value such as 0.01. (4) The first to d-th columns of R are matrix A, and the d+1-th to d+M-th columns are matrix B.

[0043] Next, a method for the user 380 to specify a data range will be described with reference to FIG. 5 . FIG. 5 is an example of a screen 550 displayed by the input / output unit 410 during model construction. The left frame of the screen 550 is an area showing actual measurement data 551 measured according to the procedure shown in FIG. 3 , and the right frame is an area showing pressure loss confirmation 570 for a data interval 564 of the actual measurement data specified by the user 380. A graph 552 showing the measured pressure values ​​of each component is displayed above the frame for the actual measurement data 551. The horizontal axis of the graph 552 represents the elapsed time (unit: seconds), and the vertical axis represents the pressure value (unit: MPa) at the measurement point measured by the pressure gauge. A solid line 553 represents the combined discharge pressure of air compressors 2 to 4, a dotted line 554 represents the pressure value at terminal 1, and dashed lines 555 and 556 represent the pressure values ​​at terminals 2 and 3. Here, the first to third ends of Zone 1 (e.g., corresponding to ends 11 to 13 in Figure 1) are shown as an example, but as shown in Figure 1, there are many ends, so the user can specify the end they want to check and display screen 550.

[0044] A graph 560 showing the measured flow rate values ​​at each point is displayed near the center of the frame of the actual measurement data 551. The horizontal axis of the graph 560 represents the passage of time (unit: seconds), which is aligned with the horizontal axis of the graph 552. The vertical axis of the graph 560 represents the flow rate value (unit: 10 －２ The flow rate (k / s) is shown in the graph 552. The dotted line 561 represents the flow rate value at Terminal 1, and the dashed lines 562 and 563 represent the flow rates at Terminals 2 and 3. It can be seen that when compressed air is consumed at specific times as shown in graph 560 at each of Terminals 1 through 3, changes in flow rate values ​​561 through 563 are observed. This consumption of compressed air causes a temporary drop in each of the pressure values ​​554 through 556, as shown in graph 552. Even though the change in flow rate values ​​561 through 563 is the same for Terminals 1 through 3, the pressure fluctuation values ​​vary from one terminal to the other, as shown in graphs 554 through 556. This is because the shape, diameter (or cross-sectional area), and length of the conduits from the discharge ports of Air Compressors 2 through 4 to the terminals are different. By displaying the pressure and flow rate data for Terminals 1 through 3 side-by-side on a time axis in this way, the user 380 can confirm whether or not the pressure drops when flow is generated (i.e., when the valve opens), and at what timing and in what manner.

[0045] Below the box for actual measurement data 551, the user 380 specifies data interval 564 and enters the prediction point number L and delay point number d in fields 565 and 566. After entering data in fields 564 through 566, the user 380 selects the model construction button 567 to construct a terminal pressure prediction model. The pressure loss at each terminal predicted by the constructed terminal pressure prediction model is displayed in graph 571 in the box for pressure loss confirmation 570. As shown in legend 572, "◯" indicates Terminal 1, "△" indicates Terminal 2, and "X" indicates Terminal 3. In this way, the difference between the discharge pressure (or tank pressure) and terminal pressure at Terminals 1 through 3 is plotted against the flow rate. Furthermore, theoretical values ​​582 through 584 of internal pipe friction loss for three typical circular pipe shapes (diameter, length) are also plotted with dotted lines, allowing the user to confirm that the differential pressures for the theoretical values ​​582 through 584 are appropriate (on the same order as the internal pipe friction). The theoretical value can be calculated by the Darcy-Weisbach formula if the flow rate, diameter, and length at the end are known.

[0046] Figure 6 shows the system configuration when operating actual equipment using an end pressure prediction model for all terminal measurements according to an embodiment of the present invention. The difference compared to Figure 2 is that the pressure gauges (113, 123, ..., 243 in Figure 2) installed at all terminals are removed, and compressed air-using equipment is connected instead of the valves (111, 121, 241 in Figure 2). Flow meters 112, 122, ..., 242 remain connected. The end pressure prediction control system 400 can continue to use the same equipment (e.g., a personal computer) as shown in Figure 2. In other words, the end pressure prediction control system 400 can be used not only to build an end pressure prediction model, but also as a system for controlling the operation of air compressors in actual pneumatic equipment using the built end pressure prediction model. When the terminal pressure prediction control system 400 is constructed on a personal computer, it may be realized by additionally installing a computer program that realizes the functions of the terminal pressure prediction control system 400 on a computer used to control factory equipment, or it may be configured so that the terminal pressure prediction control system 400 is installed and executed on another server. Furthermore, the terminal pressure prediction control system 400 for constructing a terminal pressure prediction model and a system that controls the operation of air compressors in actual pneumatic equipment using the constructed terminal pressure prediction model may be constructed on separate PCs or the like.

[0047] 6 is in operation, the outputs of the pressure gauges 305, 306, ..., 307 installed in the air compressors 2, 3, ..., 4 and the outputs of the flow meters 112, 122, ..., 242 installed in the pipes 11, 12, ..., 24 are transmitted to a terminal pressure prediction unit 402 of the terminal pressure prediction control system 400 via a communication network 360. The communication network 360 may be realized by a direct connection using a cable, or by one-way or two-way wired or wireless communication using an existing network.

[0048] When the devices 115, 125, ..., 245, etc. are actually operating, an operating unit number / rotation speed operation unit 404 of the terminal pressure prediction control system 400 controls the operation of the multiple air compressors 2, 3, ..., 4. An terminal pressure prediction unit 402 acquires in real time the outputs of the pressure gauges 305, 306, 307 and the outputs of the flow meters 112, 122, ..., 242 via the communication network 360. The terminal pressure prediction unit 402 predicts the pressure supplied to the devices 115, 125, ..., 245 using the terminal pressure prediction model created in the procedures of Figures 3 and 4 and the actually measured values ​​of the flow rates at each terminal, and a set pressure search unit 403 determines the discharge pressure (set pressure) of the air compressor supplied to the piping network 8 using the predicted value. The determined set pressure is sent to the number of operating units / rotation speed operation unit 404, which selects the air compressors 2, 3, ..., 4 to be operated and sets the operating rotation speed of the motors of the selected air compressors. A control signal from the number of operating units / rotation speed operation unit 404 is transmitted to the air compressors 2, 3, ..., 4 via a control signal line 405 or by wired or wireless communication using a communication network (not shown), and the air compressors 2, 3, ..., 4 operate in accordance with the instructions of the number of operating units / rotation speed operation unit 404.

[0049] FIG. 7 is a diagram showing an example of an input / output screen 470 in the input / output unit 410 of FIG. 6 . This screen accepts input information (control parameters, target pressure) from the user 80, and the terminal pressure prediction unit 402 displays the change in pressure (actual values ​​and future predicted values) at the terminals 11 and 21 when control is performed in accordance with the accepted input information. The left frame of the screen 470 displays input fields for a control parameter input unit 471 and a target pressure input unit 475. The user 80 inputs a control period 472 and a prediction score 473 into the control parameter input unit 471 using a keyboard or the like. The control period 472 is the period for controlling the discharge pressure from the air compressors 2, 3, ..., 4, and the prediction score 473 is a parameter indicating how far into the future pressure fluctuations are expected to be controlled from the present time. The target pressure input unit 475 inputs pressure values ​​476 and 477 as the lower limits of the terminals 11 and 21. When input into the control parameter input section 471 and the target pressure input section 475 is completed, the user 80 clicks on the control start icon 478, whereby operation control of the air compressor is executed using the created terminal pressure prediction model.

[0050] A monitor screen 480 for compressed air usage is displayed on the right side of the screen 470. Here, the amount of compressed air usage measured by the flow meters installed at the terminals 11 and 21 is displayed in a graph. The horizontal axis represents the passage of time, and the vertical axis represents the amount of compressed air usage (unit: m ３ / min). The horizontal axis shows the amount of air used at the end 11 up to the current time 17 and the amount of air used at the end 21 up to the current time 27. Below the compressed air use amount monitoring screen 480, a terminal pressure control result monitoring screen 481 is displayed. In this column, the horizontal axis shows the passage of time, and the vertical axis shows the pressure value (unit: MPa).

[0051] The pressure 18 at the terminal 11 and the pressure 28 at the terminal 21 are displayed as estimated values ​​up to the current time when viewed along the horizontal axis. The pressures 18 and 28 are estimated values ​​because, when the pneumatic equipment is in operation, the pressure gauges (113, 123, ..., 243 in Figure 2) are removed from the terminals 11 and 21, and the pressures are calculated using a terminal pressure prediction model from the actual measured values ​​of the flow meters 112, 122, ..., 242. Furthermore, predicted values ​​for the pressure 18 at the terminal 11 and the pressure 28 at the terminal 21 from the current time onward are displayed as dashed lines. These predicted values ​​are calculated by the terminal pressure prediction unit 402 using the terminal pressure prediction model stored in the model DB 430. At the bottom of the compressed air usage monitoring screen 480, the discharge pressure 37 of the air compressor is displayed as both an actual value indicated by a solid line and a predicted value indicated by a dotted line. The actual value of the discharge pressure 37 is based on the measured values ​​from the pressure gauges 305-307 installed in the air compressors 2-4 shown in Figure 6. The predicted values ​​indicated by dashed lines are values ​​set by the set pressure search unit 403. As described above, during actual operation, the terminal pressure state (pressures 18, 28) and the air compressor discharge pressure 37 are displayed in real time on the display screen of the input / output unit 410, so that the air compressor can be operated appropriately in accordance with the parameters input by the user 80 and the target terminal pressure. Note that although other terminal pressures (e.g., 12 to 14, 22 to 24) are not shown in Figure 7, the terminal pressure to be displayed on the monitoring screen 481 can be selected arbitrarily.

[0052] Next, a method for estimating the optimum pressure taking multiple terminals into consideration will be described with reference to Figure 8. The set pressure search unit 403 monitors the pressure margin at multiple terminals to be measured according to the usage status of compressed air supplied via the piping network 8, predicts future demand, and controls the operation of the air compressors 2 to 4 so that the pressure at the terminal with the smallest pressure margin is equal to or higher than the lower limit. Specifically, the aim is to lower the discharge pressure of each of the air compressors 2 to 4 as much as possible, thereby operating the entire system efficiently and contributing to energy conservation. Figure 8A shows the amount of compressed air used per hour (unit: m) at terminal 11 and terminal 21 when there are two zones. ３ 8A is a graph showing the amount of compressed air used in the equipment connected to the terminal of the device at time t １ ~t４ Here, the amount of compressed air used at time t １ ~t ２ The amount of use of terminal 11 and terminal 21 is 0.5 m ３ / min and time t ２ ~t ３ In this case, the amount of use of the end 21 is 0.6 m ３ / min, and at time t ３ Again 0.5m ３ On the other hand, the terminal 11 shows the state where the time t １ ~t ３ is constant until time t ３ The amount used is 0.6 m ３ / min.

[0053] Graphs 8B and 8C show pressure fluctuations at the ends 11 and 21 when the amount of compressed air used at the ends 11 and 21 is as shown in FIG. 8A. In FIG. 8B, the typical pressure at the end 11 in zone 1 (measured by the pressure gauge 113 in FIG. 2) is １ ~t ２ The pressure is about 0.66 MPaG, which is sufficiently larger than the set target value (lower limit) of the air pressure, and is maintained at a high pressure value with a margin of about 10% from 0.6 MPaG. ２ As shown in FIG. 8A, the amount of compressed air used by the equipment connected to the terminal 21 is 0.6 m ３ / min, the discharge pressure 37 of the air compressor is controlled by the set pressure search unit 403 (see FIG. 6) to rise suddenly from arrow 37a as shown by arrow 37b in FIG. 8D, following the increase in the amount of compressed air used at the end 11. As a result of this temporary increase in discharge pressure 37, the pressure 18 at the end 11 temporarily increases as shown by arrow 18a to 18b, but in the case of the pneumatic equipment of this embodiment, the pressure 18 quickly returns to its original value.

[0054] In FIG. 8C, the pressure at the end 21 (measured by the pressure gauge 213 in FIG. 1) is １ ~t ２The pressure is about 0.63 MPaG, which is a small pressure value with a margin of about 5% from the set target value (lower limit) of the air pressure. However, when the amount of compressed air used by both the terminals 11 and 21 is less than the specified value of 0.5 m ３ / min, time t １ ~t ２ The margin value near the pressure 18d at the end 11 of the middle is smaller than the margin value near the pressure 28d at the end 21. This magnitude relationship of the margin values ​​occurs due to differences in the piping structure (flow path resistance due to piping length, thickness, bending, etc.). ２ 8A, when the amount of compressed air used on the end 21 side increases, the pressure 28 at the end 21 temporarily decreases as indicated by arrow 28b. Here, the margin value 28c near the pressure 28a at the end 21 is small (small margin) at about 5%, so the pressure drop near arrow 28b falls below the lower limit of the target pressure, 0.6 MPaG. However, due to the effect of the control to increase the discharge pressure 37 indicated by arrow 37b in FIG. 8D, the pressure quickly returns to its original value.

[0055] Time t ３ In this case, when the compressed air usage amounts 17, 27 shown in Fig. 8A increase on the end 11 side and return to the original usage amount on the end 21 side, as shown in Fig. 8D, the discharge pressure 37 of the air compressor in Fig. 8D is increased as shown by arrow 37d under the control of the set pressure search unit 403 (see Fig. 2), and is controlled to follow the increase in usage amount at the end 11. At this time, although the pressure at the end 11 temporarily decreases as shown by arrow 18c in Fig. 8B, since the margin value of the pressure 18 is set to a large value, even if the pressure decreases as shown by arrow 18c, it remains at or above the lower limit value of the target pressure of 0.6 MPaG. Due to the effect of the control to increase the discharge pressure 37 as shown by arrow 37d, as shown in Figs. 8B and 8C, ３ After a short time the pressure at the ends 11, 21 will maintain its original value.

[0056] As shown in Figures 8A to 8D, the terminal pressure prediction control system 400 uses the terminal pressure prediction model constructed by the terminal pressure prediction model construction unit 401 to monitor fluctuations in the usage status of compressed air at representative terminals of each zone, and the set pressure search unit 403 (see Figure 6) monitors and predicts the pressure margins at multiple terminals according to fluctuations in usage, and the operating number / rotation speed operation unit 404 sets the number of operating units and rotation speeds of the multiple air compressors 2 to 4 so that the pressure at the terminal with the smallest pressure margin is equal to or greater than the lower limit. As a result, as shown in Figure 8D, the set pressure search unit 403 (see Figure 6) １ ~t ２ Section A up to time t ２ ~t ３ In the section B up to time t, the end 21 is the main control target so that the pressure value of the end 21 does not fall below the lower limit. ３ In the subsequent section C, the terminal 11 is the main control target so that the pressure value at the terminal 11 does not fall below the lower limit. In this way, in Example 2, the user 380 inputs the specifications (lower limit pressure, rated air flow rate, frequency of use) of each device connected to the terminal, determines a representative terminal to be measured, and estimates the time change pattern of the terminal pressure based on the measured flow rate of the representative terminal and the terminal pressure prediction model. This makes it possible to use the terminal pressure prediction model to lower the discharge pressure at each terminal as much as possible depending on the compressed air usage situation. As a result, an air compressor control system that reduces unnecessary power consumption and promotes energy conservation has been realized.

[0057] 9 is a diagram showing an example of control for suppressing hunting due to a delay in the terminal pressure fluctuation in the terminal pressure predictive control system 400 of FIG. 2. The boxed diagram on the left shows the pressure fluctuation situation when a conventional control method, i.e., a method that does not perform terminal pressure predictive control using a terminal pressure predictive model, is performed and the terminal instantaneous value is fed back (FB) without considering the transmission delay of the terminal pressure fluctuation. The boxed diagram on the right shows the pressure fluctuation situation when terminal pressure predictive control is performed using a terminal pressure predictive model according to an embodiment of the present invention. These controlled objects are for the pneumatic system shown in FIG. 1, with zones 1 and 2 defined, and zone 1 showing the pressure fluctuation at terminal 11 and zone 21 showing the pressure fluctuation at terminal 21.

[0058] The top graph in the box on the left and right shows the amount of compressed air used (amount of compressed air used) at the end 11 and the end 21 (unit: m ３ / min) 171, 175. The horizontal axis is time. In both the left and right graphs, the time t １ ~t ６ In this case, the compressed air consumption 171, 175 at the end 11 is ２ The amount used is 0.5 m ３ / min to 0.6m ３ / min, while the compressed air consumption 271, 275 at the end 21 increases at time t ２ From t ６ Usage up to 0.45m ３ The second and third graphs in the boxes on the left and right show the pressure at the end 11 and the pressure at the end 21 when the compressed air usage amounts 171 and 175 change.

[0059] In the box on the left showing the conventional example, at time t ２ When the amount of compressed air at the end 11 increases by about 20%, as shown in the second graph, the pressure of the compressed air at the end 11 increases at time t ２ ~t ３ In the figure, the pressure at end 21 on the zone 2 side decreases slightly as indicated by arrow 280a. This phenomenon does not directly correspond to an increase in the amount of compressed air used 171 at end 11, but occurs with a delay relative to the change in the flow rate of compressed air. Similarly, the pressure at zone 1 where end 11 is provided decreases, and as a result, the pressure at end 21 on the zone 2 side also decreases slightly as indicated by arrow 280a. The timing and magnitude of the effect of the decrease in pressure at end 11 on end 21 are greatly affected by the diameter and length of mother pipes 10 and 20, so what is shown here is merely an example.

[0060] In this way, time t ２ ~t ３ When a pressure decrease occurs at time t, the pressure value is fed back (FB) to a conventional terminal pressure control system (not shown), and the terminal pressure control system controls the discharge pressure 370 of the air compressor as shown by arrow 370a at time t ２ From t ３By this control of increasing the discharge pressure 370, the pressure at the end 11 increases with a time delay as shown by the arrow 180b, up to time t ３ ~t ４ Since the header 20 on the end 21 side is connected to the header 10 on the end 11 side, the pressure at the end 21 decreases as shown by the arrow 280a, and then increases with a time delay as shown by the arrow 280b at time t ３ ~t ４ It increases slightly at

[0061] Time t ４ At time t, the pressure at the end 11 recovers to the set pressure as indicated by the arrow 181. This recovery occurs when the discharge pressure 370 on the air compressor side reaches the set pressure. ３ In other words, the increased pressure of the arrow 371 is transmitted to the end 11. ４ -t ３ The change in discharge pressure 370 is transmitted to the terminal 11 with a time lag of 180c. In feedback control, the presence of a transmission delay with respect to such a change in discharge pressure 370 leads to the occurrence of a hunting phenomenon in which the pressure repeatedly increases and decreases, as shown by arrows 180c and 180d. Furthermore, because the occurrence of the hunting phenomenon at the terminal 11 is fed back, the terminal pressure control system attempts to control the discharge pressure of the air compressor to vary so as to cancel out the hunting phenomenon, and as a result, the discharge pressure 370 of the air compressor is increased or decreased as shown by arrows 370b to 370d, resulting in a hunting phenomenon 372. The occurrence of a hunting phenomenon due to a delay in the change in terminal pressure not only leads to energy loss, but also results in unnecessary power consumption by the air compressor, which is undesirable from the perspective of energy conservation.

[0062] The graph in the box on the right shows the control by the terminal pressure predictive control according to the embodiment of the present invention. ２ In this case, the amount of compressed air used on the terminal 11 side is 0.5 m ３ / min to 0.6m ３ / min, the terminal pressure prediction unit 402 (see FIG. 6) predicts the pressure 185 at the terminal 11 and the pressure 285 at the terminal 21 for a certain period in the future. In this prediction, (a) the terminal pressure prediction unit 402 predicts the future values ​​of the terminal pressures 185 and 285 at the terminals 11 and 21 when the discharge pressure 375 of the air compressor is set to various pattern candidates, (b) searches for a prediction pattern that minimizes the cumulative error from the target value of the predicted terminal pressure, and (c) adopts the prediction pattern that minimizes the cumulative error determined by the search and predicts the future values ​​of the terminal pressures 185 and 285 for the terminal 11 and 21 within the prediction period (here, time t ２ ~t ４ ) air compressor operation.

[0063] In the bottom graph in the box on the right, time t ２ In this example, the terminal pressure prediction unit 402 prepares multiple change patterns for the discharge pressure 375 on the air compressor 2-4 side. Three patterns are derived: discharge pressure 375a (first candidate) indicated by a thick solid line, discharge pressure 375b (second candidate) indicated by a dashed line, and discharge pressure 375c (third candidate) indicated by a dotted line. Once these three patterns are calculated, the set pressure search unit 403 (see FIG. 6) determines which of these candidates is most suitable.

[0064] When the first to third change candidates 375a to 375c for the discharge pressure 375 are set in this way, the terminal pressure prediction unit 402 uses the terminal pressure prediction model stored in the model DB 430 (see FIG. 6) to predict the terminal pressure at the time t ２ ~t ４The pressure fluctuations leading to the terminal pressure 185a are predicted. As a result of this prediction, three patterns are calculated for the terminal 11: pressure 185a (first candidate) shown by a thick solid line, pressure 185b (second candidate) shown by a dashed line, and pressure 185c (third candidate) shown by a dotted line. Similarly, for the terminal 21, pressure 285a (first candidate) shown by a thick solid line and pressure 285b (second candidate) shown by a dashed line are calculated. Note that the third candidate 285c (third candidate), which should be shown by a dotted line, is the same as the second candidate and is therefore shown by the same dotted line. Considering these candidates and the pressure fluctuations at the terminals 11 and 21, it can be determined that the air compressor should be controlled at the first discharge pressure 375a, which has the least pressure fluctuation and is most preferable. The terminal pressure prediction unit 402 uses the first candidate pressures 185a, 285a, and discharge pressure 375a, and by displaying these graphs with thick solid lines, it is possible to clearly indicate to the user 80 viewing the screen which candidate is being used.

[0065] As described above, the terminal pressure predictive control system 400 searches for the optimal setting value for the discharge pressure of the air compressor using the terminal pressure predictive model, and controls the operation of the air compressor at the discharge pressure 375a determined by the search. If multiple air compressors are provided on the high-pressure air supply side, the number of operating units / rotation speed operation unit 404 (see Figure 6) of the terminal pressure predictive control system 400 controls the on / off of each air compressor and even the rotation speed of the motor that drives the compressor during operation.

[0066] In the right diagram of FIG. ２ The terminal pressure prediction unit 402 predicts the pressure at the terminals 11 and 21 during the prediction period. ４ However, such a prediction is made up to the end of time t ２ Instead of just doing it once at time t ２ Thereafter, the process continues at predetermined time intervals, and the optimum discharge pressure of the air compressor at each execution is sequentially set. ２ The interval between the next predictive control and the interval between the prediction start point and the prediction end point in each predictive control can be set arbitrarily, and the optimal time interval can be set taking into consideration the characteristics of the pneumatic equipment in the factory or work site.

[0067] Figure 10 is a flowchart of compressor control in the terminal pressure prediction control system 400 of Figure 6. The flowcharts of Figure 10 and Figure 11 (described later) are executed by the set pressure search unit 403 and the number of operating units / rotation speed operation unit 404, and are realized by a processor (not shown) included in the terminal pressure prediction control system 400 executing a specific computer program. First, the set pressure search unit 403 acquires a terminal pressure prediction model from the model DB 430 in which the created terminal pressure prediction model is stored (step 521). Next, the set pressure search unit 403 executes multiple steps 522 to 526, enclosed by a dashed line 527, at regular intervals. Steps 522 to 526 are closed-loop control processing executed at regular intervals.

[0068] In the closed-loop control portion 527, the set pressure search portion 403 acquires the discharge pressure history and terminal flow rate value of each of the air compressors 2 to 4 from an operation history database (not shown) (step 522). Next, the set pressure search portion 403 searches for the optimum value of the discharge pressure of the air compressors 2 to 4 in the next control cycle (step 523). Here, the detailed procedure of step 523 (subroutine for searching the optimum value of the discharge pressure) will be described using the flowchart in FIG. 11.

[0069] 11, first, the set pressure search unit 403 sets a candidate value for the discharge pressure in the next control cycle (step 531). Next, the set pressure search unit 403 predicts the terminal pressure until H cycles later when the candidate value is held from the current time until H cycles later (step 532). This prediction is made by dividing the current time by t ０ Then, at time t=t ０ +Δt ｃ From t = t ０ +HΔt ｃ Up to the compressor discharge pressure P ｃｏｍｐ (t) to P ｃｏｍｐ (t ０ ) + ΔP ｃｏｍｐ This is the end pressure when the pressure is maintained at Δt ｃ (≧Δt) is the control period.

[0070] Next, the set pressure search unit 403 calculates the cumulative error of the terminal pressure up to the predicted H period (step 533). Next, the set pressure search unit 403 calculates an evaluation function (step 534). This calculates the cumulative error and evaluation function according to the following formula:

[0071] Here, the whole of ( ) is the evaluation function, and the part in [ ] of ( ) is the cumulative error. In the formula [ ], the negative column in front is the predicted value of the pressure at M terminals, and the back r １ ~r Ｍ The column indicates the target pressure values ​​of the terminals (M terminals).

[0072] Next, it is determined whether the evaluation function shown in Equation 9 has converged (step 535). If the evaluation function has not converged, the candidate value for the discharge pressure is corrected and the process returns to step 531 (steps 535, 537). If the evaluation function has converged in step 535, the set candidate value is used as the discharge pressure in the next control cycle (step 536). This ΔP ｃｏｍｐ The compressor discharge pressure is calculated using the search value according to the following equation 10.

[0073] Next, returning to FIG.

[0074] The value found by the subroutine of Figure 11 is set as the discharge pressure of air compressors 2-4 and sent to the number of operating units / rotation speed operation unit 404 (step 524). The number of operating units / rotation speed operation unit 404 operates the number of operating air compressors and their respective rotation speeds so that air compressors 2-4 are operated based on the specified value (step 525). Next, the processor of terminal pressure prediction control system 400 determines whether or not to stop the control shown in Figure 10 (step 526). If not, the processor returns to step 522 at the start of the next control cycle and repeats steps 522-526. If the processor determines in step 526 that the control of Figure 10 should be stopped, the control is terminated.

[0075] As described above, a typical end-of-service flow rate and actual measured values ​​of the air compressor discharge pressure are input into the end-of-service pressure prediction model. This computer program then predicts the typical end-of-service pressure, enabling the search for the optimal set pressure of the air compressor so that the predicted representative end-of-service pressure approaches the target pressure set by the user. This control allows for efficient control of the number of operating air compressors and their rotation speeds. Furthermore, optimizing the future values ​​of pressure at multiple end-of-service pressures can suppress the hunting phenomenon 372 shown in Figure 9 and achieve control tailored to the end 21 with the least pressure margin. In this way, measuring actual air consumption at all end-of-service pressures and estimating the optimal pressure taking all end-of-service pressures into account enables optimal air compressor operation control and effectively suppresses hunting. This process is modeled after model predictive control (MPC) and can be easily implemented using known OSS programs, libraries, or commercial technical calculation software.

[0076] Next, a second embodiment of the present invention will be described using Figures 12 to 15. In the first embodiment, the actual air consumption at all terminals is measured and a terminal pressure prediction model is created. During actual operation, optimal air compressor operation control is performed by using the actual measured values ​​of air consumption at all or some of the terminals and the terminal pressure prediction model. However, because pneumatic equipment in a factory can range from small to large, and the number of pipes can range from tens to hundreds, it can be difficult to measure the pressure values ​​and air flow rates at all terminals to create a terminal pressure prediction model. Therefore, in the case of large-scale pneumatic equipment, the piping pressure of the pneumatic equipment is divided into multiple zones, and the pressure and air consumption at a representative terminal in each zone are measured to create a terminal pressure prediction model for each zone. During actual operation of the pneumatic equipment, a pressure gauge is removed from a representative terminal (note: it can remain attached), and the air consumption at each terminal is estimated. This is input into the terminal pressure prediction model to predict the terminal pressure value. Based on this predicted value, the set pressure search unit 403 sets the optimal air compressor discharge pressure.

[0077] FIG. 12 is a diagram illustrating a model operation configuration during zoning in a terminal pressure predictive control system 400 according to a second embodiment of the present invention. A user 380 selects one "representative" pipe from each zone, installs pressure gauges 113, 363, etc. at the terminal of the representative pipe to measure pressure values, and installs flow meters 112, 362, etc. to measure pressure values ​​and air consumption. These measurements and the outputs of pressure gauges 305-307 are used to create a terminal pressure predictive model for the selected terminal. During preliminary measurements for creating the terminal pressure predictive model, the user 380 divides the entire piping network into N zones (N > 1) based on the physical characteristics of the piping network, such as the branching of the main pipe, the shape of the piping (length, diameter, bends, etc.), and the presence or absence of loops. It is preferable that each zone includes one or several terminals. The user 380 who divides the network into zones 1 to N may be a user of the pneumatic equipment or a service technician responsible for the installation and operation of the terminal pressure predictive control system 400. Zoning is performed manually in this manner, and the zoning definition and representative point information are stored in the terminal definition database (DB) 440 by the user 380 .

[0078] In Fig. 12, the equipment upstream from the piping network 8, i.e., air compressors 2, 3, ..., 4, piping 302, 303, ..., 304 to the piping network 8, and pressure gauges 305, 306, ..., 307 that measure air pressure, are configured in the same manner as shown in Fig. 2. A plurality of terminal pipes 11...14, 31-36 branch off from the piping network 8. Here, the illustration of the downstream portion from the piping network 8 in terms of the flow of compressed air differs from Fig. 2, and only the piping portion corresponding to zone N is shown, but the configuration of the pneumatic equipment in zones 1 and 2 is the same as in the first embodiment.

[0079] The operator defines the terminals 11 to 14 of the multiple terminals 11...14, 31 to 36 connected to the piping network 8 as zone 1, and defines terminal 11 as the measurement target. These terminals 11 to 14 correspond to terminals 11 to 14 in the piping configuration shown in Figure 1. Similarly, the operator zones the multiple piping groups from zone 2 onwards. In zone N, not shown in Figure 1, terminals 31 to 36 exist, and terminal 36 is designated as the control target, and actual values ​​of flow rate and discharge pressure are measured. To perform these measurements, the user 380 removes the high-pressure air-using equipment (not shown) that receives the compressed air from the representative terminals (terminals 11 and 36 in this case) and installs valves 111 and 361 instead. Furthermore, flow meters 112 and 362 and pressure gauges 113 and 363 are installed in the pipelines between terminals 11 and 36. A representative terminal is also selected from zones 2 to N-1 (not shown), and a valve, flow meter, and pressure gauge are installed at the selected terminal. At the non-representative ends of zones 1 to N, no flow meter or pressure meter is separately connected, and the existing valve (not shown) is kept closed to prevent compressed air from coming from the ends.

[0080] The user 380 operates the valves, and the outputs of the flow meters 112, 362 and the pressure meters 113, 363 at each representative terminal at that time are recorded as actual measured values ​​of flow rate and discharge pressure with the acquisition times synchronized in an experimental database (DB) 420 of the terminal pressure prediction control system 400. Note that if acquisition of actual measured values ​​of flow rate and discharge pressure at all terminals in the first embodiment has been completed before constructing the model operation during zoning in the second embodiment, a terminal pressure prediction model for each representative terminal may be constructed using the actual measured values ​​of all terminals.

[0081] The model construction procedure after the measurement data is recorded in the experimental database 420 is basically the same as the procedure in the first embodiment. By setting one representative terminal (or two or more, but less than the total number of terminals) as in the second embodiment, it becomes easy to apply the air compressor pressure control system according to the present invention to pneumatic equipment in facilities such as large-scale factories where there are hundreds of terminal pipelines.

[0082] Next, a method for defining terminal information will be described using the terminal management table 450 in Fig. 13. First, during measurements before building a model, the user 380 defines each zone and a representative terminal (representative point) for each zone. Specifically, by regarding the area from the branch 310 of the main pipe 300 to the ends of multiple terminal pipelines as one "zone," the user defines "zones" from 1 to N (where N is a positive natural number), sets one representative terminal to be measured for each zone, and measures the pressure and flow rate at the set representative terminal.

[0083] The terminal management table 450 in FIG. 13 is stored in advance in a terminal definition database (DB) 440 defined in a storage device (not shown) of the terminal pressure predictive control system 100, and defines the model operation configuration during zoning in the terminal pressure predictive control system 400. The terminal management table 450 is broadly divided into storage areas for a terminal definition 451 and a flow rate information definition 461 for each terminal. The terminal definition 451 stores detailed information for each terminal, uniquely defined by a terminal number 453. Here, the example illustrates zones 1 and 2 corresponding to FIG. 1 and their respective terminal numbers 11 to 14 and 21 to 24. The representative point 454 is a representative point designated by the user 380, and one representative point 454 is selected for each zone. In the configurations of FIGS. 1 and 7, terminal number 11 is selected for zone 1, and terminal number 21 is selected for zone 2. A check mark is placed in the representative point 454 column to indicate that the representative point is a representative point. Furthermore, additional information for identifying the representative point 454, such as location information 455, is added to the representative terminals 11 and 21. Image information such as text, coordinate information, marks on a piping diagram, and photographs is stored as the location information 455. In this example, the terminal with terminal number 453 "11" has a location description recorded in text information, such as "the end of the straight pipe extending directly from the compressor." Meanwhile, the terminal with terminal number 453 "21" has stored, as location information 455, the file name "zone2.jpg" of image data of the terminal position photographed by a still image camera.

[0084] The flow rate information definition 461 for each terminal stores the rated airflow rate 462 and operation method 463 of the equipment connected to each terminal. The user 380 inputs the rated airflow rate 462, operation method 463, availability rate 464, and frequency 465 of the equipment used at each terminal. The rated airflow rate 462 is data displayed on a product nameplate or label, etc., which displays the model number and specifications of the connected equipment. The operation method 463 indicates whether the operation is "continuous," which consumes compressed air continuously, or "intermittent," which consumes compressed air intermittently. The availability rate 464 indicates the ratio of operation time to downtime in the case of intermittent operation. The availability rate 464 can be obtained from a product nameplate or label, instruction manual, etc., or by acquiring actual operation data. Here, the availability rate of the equipment connected to terminal number "23" is 80%, which can be calculated by tallying the total operation time and the total downtime within a specified cycle time and calculating the ratio between them. The operation rate may be automatically acquired from monitoring data of a device using compressed air (not shown) and input as the operation rate. The frequency 465 stores the occurrence timing (cycle) of intermittent operation.

[0085] Once the zone definition information and the definition of the devices connected to the terminals are completed in the terminal management table 450, the terminal pressure prediction model construction unit 401 expands the constructed terminal pressure prediction model based on the zone definition information so that it can also predict non-representative points. Specifically, from matrix A and matrix B of the terminal pressure prediction model constructed for N representative points, components (submatrices) related to representative points belonging to the same zone as the non-representative point to be calculated are extracted and used as the terminal pressure prediction model for that non-representative point. The method for constructing the terminal pressure prediction model is the same as the method shown in the flowcharts of Figures 3 and 4, except that the number of terminals to be measured is reduced. The model formula used here can be calculated using the following formula 11.

[0086] Here, Pcomp(t), ..., Pcomp(t-d) are the discharge pressures of the air compressor at time t, ..., time t-d. １ ２ (t), U ２ ２(t) is the square of the flow rate at the terminal 11 (a representative terminal of zone 1) and the terminal 21 (a representative terminal of zone 2) at time t, and Δt is the sampling rate (assumed to be about 1 second). １，１ ... ] is the matrix coefficient A, and the brackets [b １，１ …b ２Ｌ，２ ] is the matrix coefficient B.

[0087] Then, using the created terminal pressure prediction model, a terminal pressure prediction unit 402 (see FIG. 12) calculates an estimated value of a representative terminal pressure, and a set pressure search unit 403 determines the discharge pressure of the air compressor that is in line with the estimated pressure value. An operating unit number / rotation speed operation unit (operation operation unit) 404 actually operates air compressors 2 to 4 in accordance with the discharge pressure determined by the set pressure search unit 403. In this way, optimal air compressor operation control is possible by measuring the actual air usage at a representative terminal in each zone and estimating the optimal pressure taking all terminals into consideration.

[0088] As described above, in the second embodiment, the area from the branch point 310 (see FIG. 1) of the parent pipe 300 to the end was considered a single "zone," and the pressure and flow rate of one end for each zone were measured to create an end pressure prediction model for those zones. As an extension of the second embodiment, it is possible to create an end pressure prediction model that includes ends that were not measured at the time of model creation from the end pressure prediction model created in the second embodiment. In other words, using trained parameters, an extended end pressure prediction model that includes ends that were not measured at the time of model creation can be created. The scope of this extended end pressure prediction model is as shown in the schematic configuration diagram in FIG. 14, with ends 1a to 1c of zone 1 included and ends 2a to 2c of zone 2 included.

[0089] FIG. 15 shows the equation for the terminal pressure prediction model created in this modified example. Here, as shown in FIG. 12, pressure meters and flow meters are installed at representative terminals (e.g., terminals 1a and 2a), and pressure and flow data are measured as in Example 1. Using this data, terminal pressure prediction models for terminals 1a and 2a are constructed as in Example 1. In this model, the pressure fluctuation characteristics of the terminals within each zone are considered to be the same, and a terminal pressure prediction model including terminals that were not measured at the time of construction is constructed using learned model parameters, as in the equation of FIG. 15. Reference numeral 52 denotes matrix coefficient A of the terminal pressure prediction model, and reference numeral 54 denotes matrix coefficient B. These matrix coefficients A and B correspond to the matrix coefficients shown in Equation 1.

[0090] Here, the meaning of the coefficients of the end pressure prediction model shown in Fig. 15 is as follows: Coefficient a including the parts indicated by dotted lines 52a and 52d in matrix coefficient A (arrow 52) ２ｍ－１，１ ,... ,a ２ｍ－１，ｄ (m=1, 2, . . . , L) is a coefficient that represents the time evolution of the pressure in zone 1, and is a coefficient a including the parts indicated by the dotted lines 52b and 52e. ２ｍ，１ ,... ,a ２ｍ，ｄ (m=1, 2, . . . , L) is a coefficient that represents the time evolution of the pressure in zone 2. The coefficient b (arrow 54) of the matrix coefficient B includes the part surrounded by the dotted line 54a or the dotted line 54d and the dotted line 54f. ２ｍ－１，１ ,... ,b ２ｍ－１，１ (m=1, 2, . . . , L) is a coefficient that represents the influence of the use of compressed air in zone 1. The coefficient b that includes the part of the matrix coefficient B that is surrounded by the dotted line 54b or the dotted line 54e and the dotted line 54g ２ｍ，２ ,... ,b ２ｍ，２ (m=1, 2, . . . , L) is a coefficient that represents the influence of the use of compressed air in zone 2. Coefficient b including the part of matrix coefficient B surrounded by dotted line 54a or dotted line 54d and dotted line 54g ２ｍ－１，２ ,... ,b ２ｍ－１，２ (m=1, 2, . . . , L) is a coefficient that represents the influence of the use of compressed air in Zone 2 on Zone 1. The coefficient b that includes the part of the matrix coefficient B that is surrounded by the dotted line 54b or the dotted line 54e and the dotted line 54f ２ｍ，１,... ,b ２ｍ，１ (m=1, 2, . . . , L) is a coefficient that represents the effect that the use of compressed air in Zone 1 has on Zone 2.

[0091] 15 , for matrix coefficient A (first term on the right side) indicated by reference numeral 52, if the terminal of the prediction target is zone 1, a coefficient representing the time evolution of pressure in zone 1 is used, and if the terminal is zone 2, a coefficient representing the time evolution of pressure in zone 2 is used. For matrix coefficient B (second term on the right side) indicated by reference numeral 54, if the terminal of the prediction target is zone 1 and the terminal using compressed air is zone 1, a coefficient representing the influence of compressed air use in zone 1 is used. If the terminal of the prediction target is zone 2 and the terminal using compressed air is zone 2, a coefficient representing the influence of compressed air use in zone 2 on zone 1 is used. If the terminal of the prediction target is zone 1 and the terminal using compressed air is zone 2, a coefficient representing the influence of compressed air use in zone 2 on zone 1 is used. If the terminal of the prediction target is zone 2 and the terminal using compressed air is zone 1, a coefficient representing the influence of compressed air use in zone 1 on zone 2 is used.

[0092] In this way, by modeling the time evolution of pressure and the characteristics of the impact of compressed air use for each zone based on the measurement data for each zone and then using this model to model other ends within the zone, it is possible to model the pressure transmission characteristics of large-scale piping with little effort.

[0093] FIG. 16 is a table summarizing the measurement points and operating characteristics in three usage modes of the compressor control system according to the present invention. The first one, indicated by the number 601, is the compressor control system 1 described in the first embodiment. In the first embodiment, all terminals are, in principle, the measurement targets as measurement points 603 when building the model. Furthermore, as for flow rate information 604 for each terminal during actual operation after model creation, the compressor is controlled using measurement values ​​acquired at all or some of the terminals. As a result, hunting suppression is achieved and an optimum pressure estimate is obtained as output 605.

[0094] The second, designated by the reference numeral 601, is the compressor control system 1 described in the second embodiment. In the second embodiment, the terminal group is divided into multiple zones as measurement points 603 when constructing the model, and a representative terminal of each zone is selected as the measurement target, and the representative terminal is used to measure the flow rate during operation. Flow rate information 604 for each terminal during actual operation after the model is created is estimated from the use of all equipment entered by the user, without using a flow meter, to control the compressor. As in the first embodiment, hunting suppression and an estimate of the optimal pressure are obtained as output 605.

[0095] The third example, designated by the reference numeral 601, relates to Example 3, which will be described below. In Example 3, the compressor control system 1 is not used for directly controlling the multiple air compressors 2 to 4, but is used as an "energy-saving diagnostic system" for determining energy conservation in the operation of the multiple air compressors 2 to 4. In the energy-saving diagnostic system, for example, the measurement points 603 during model construction can be all terminals, but it is also possible to measure only representative points. Furthermore, flow rate information 604 for each terminal during operation (diagnosis) can be obtained from all or some of the terminals to diagnose whether the operation settings of the air compressors are appropriate. In this energy-saving diagnostic system, various calculations are performed on the customer's computer or the service provider's server, and estimated optimal operating conditions and optimal pressures are output to the customer's computer or a mobile information terminal.

[0096] FIG. 17 is a configuration diagram of a compressor set pressure recommendation system according to a third embodiment of the present invention. The basic configuration is substantially the same as that of FIG. 6, and the same reference numerals are used to designate the same components. The compressor set pressure recommendation system 400A measures the flow rates of terminals 11, 12, ..., 24, etc., in the pneumatic system of a customer factory to diagnose whether the settings for the operation control of multiple compressors 2-4, which are manually performed by a user 80, are appropriate, and displays the optimal recommended setting value (recommended set pressure) on the input / output unit 410. The user 80 inputs the operation parameters (number of compressors, discharge pressure) of the air compressors 2-4 via the input / output unit 410, and the set pressure search unit 403 displays the created end pressure prediction model, and the air compressor setting conditions (set pressure, etc.) are displayed on the input / output unit 410. The user 80, such as a production engineering department employee at the customer factory, views the display results on the input / output unit 410 and changes the setting value on the compressor number control panel 390. The compressor number control panel 390 is a device for controlling the number of operating air compressors 2 to 4, the set pressure, and other settings.

[0097] FIG. 18 shows an example of a display screen 670 in the input / output unit 410 when recommending compressor settings. Among the displayed contents, the graph in the compressed air usage amount 680 column is the same as that shown in FIG. 8A. The graphs in the discharge pressure / terminal pressure simulation screen 681 are the same as those shown in FIGS. 8B to 8D. A target pressure input section 671 in the upper left of the display screen 670 is provided with a column 672 for inputting the target pressure for terminal 11 and a column 673 for inputting the target pressure for terminal 21. While FIG. 18 shows only two columns, 672 and 673, for inputting the target pressure, it is also possible to provide a number of input columns corresponding to the number of zones or the total number of representative terminals in an actual pneumatic system.

[0098] After entering the target pressure values ​​to be set in each of fields 672 and 673, the user 80 clicks the optimization calculation icon 674. The compressor set pressure recommendation system 400A (see FIG. 17 ) then applies the terminal pressure prediction model constructed in Example 1 or Example 2 to calculate the optimal set pressures for air compressors 2 to 4. The set pressure display section 675 is a field that displays the calculated optimal values ​​676 and 677. In FIG. 18 , the current set pressure value 676 ("0.72 MPaG") and the calculated optimal value 677 ("0.62 MPaG") are displayed. This result allows the user to understand that the discharge pressure 37 of the air compressor can be reduced from the current 0.72 MPaG to 0.62 MPaG. As described above, according to Example 3, when the user 80 inputs the terminal target pressure, the set pressure that will bring the terminal pressure closest to the target value when the current usage amount is maintained constant is calculated, and this value can be displayed together with the predicted discharge pressure and terminal pressure.

[0099] Although the present invention has been described above using three embodiments, the present invention is not limited to the above-described embodiments and various modifications are possible within the scope of the invention.

[0100] 1...pressure control system, 2 to 4...air compressors, 5 to 7...pressure gauges, 8...piping network, 10, 20, 60...head pipe, 11 to 14, 21 to 24...end (end of piping), 111, 121, 131, 141...valves, 112, 122, 142...flow meter, 113, 123, 143...pressure gauge, 211, 221, 231, 241...valves, 212...flow meter, 213...pressure gauge, 300...head pipe, 302, 302, 304...piping, 305, 306, 307...pressure gauge, 310...branch, 360...communication network, 380...user, 400...terminal pressure prediction control system 401...Terminal pressure prediction model construction unit, 402...Terminal pressure prediction unit, 403...Set pressure search unit, 404...Number of operating units / rotation speed operation unit, 405...Control signal line, 410...Input / output unit, 420...Experiment database, 430...Model database, 440...Terminal definition database, 450...Terminal management table, 451...Terminal definition, 452...Zone, 453...Terminal number, 454...Representative point, 455...Information about position, 461...Flow rate information definition for each terminal, 462...Rated air volume, 463...Operation method, 464...Operation rate, 465...(Operation) frequency, 470...Display screen

Claims

1. A compressor control system for controlling equipment that supplies compressed air discharged from an air compressor through multiple pipes to multiple terminals that consume the compressed air, comprising: a pressure gauge provided at one or multiple supply points near the air compressor or a receiver tank; a pressure gauge, a flow meter, and a valve connected to all or some of the terminals of the pipes; a storage means for storing data of actual pressure and flow rate measured by opening the terminals while changing the opening degree of the valves; a terminal pressure prediction model construction unit for using the actually measured data to learn a terminal pressure prediction model having the pressure of the supply point and the flow rates of the multiple terminals as input and the terminal pressure as output; a terminal pressure prediction unit for predicting the pressure of the terminals by inputting the measured value of the pressure at the supply point during actual operation of the equipment and flow rate information for each terminal into the learned terminal pressure prediction model; and a set pressure search unit for setting an optimal discharge pressure of the air compressor using the predicted terminal pressure value.

2. The compressor control system according to claim 1, further comprising an operation control unit that controls the rotation speed and number of operating air compressors so that the multiple terminal pressures set by the set pressure search unit fall within a predetermined range.

3. The compressor control system according to claim 2, characterized in that the terminal pressure prediction unit is used to control the rotation speed and number of operating air compressors so that the future predicted values ​​of the multiple terminal pressures and the discharge pressure of the air compressor are constant.

4. The compressor control system described in claim 3, characterized in that the terminal pressure prediction model construction unit uses many of the terminal pressure measurement values ​​and flow rate information when learning the terminal pressure prediction model, and the terminal pressure prediction unit makes predictions when setting the optimal discharge pressure using a smaller number of terminal flow rate information values ​​than those used during the learning process.

5. A compressor control system as described in claim 4, characterized in that a plurality of branched portions of the piping are defined as zones, and a representative terminal end is set for each zone at which flow rate and pressure are measured, and the terminal pressure prediction unit sets the optimal discharge pressure of the air compressor using actual measurement data of the flow rate and pressure of the representative terminal end.

6. A compressor control system as described in claim 5, further comprising an input / output unit for inputting and outputting information, and a model database for storing the rated flow rate, operating method, and operating rate of the equipment connected to the end input by a user, wherein the terminal pressure prediction unit estimates the change pattern of the terminal flow rate from the input information, and predicts the measured value of the supply point pressure and the future pressure change pattern.

7. A compressor control system as described in claim 5, further comprising a display device that visually displays information to a user, and the optimal discharge pressure of the air compressor set by the set pressure search unit is displayed on the display device.

8. A compressor control system as described in claim 1, characterized in that it has an input / output unit for inputting and outputting information, and in response to a user inputting the terminal target pressure from the input / output unit, the set pressure search unit simulates the optimal discharge pressure of the air compressor and displays it on the input / output unit.

9. A compressor control method in a pneumatic system in which compressed air discharged from an air compressor is supplied to a plurality of terminals consuming the compressed air via a plurality of pipes, using a control unit having a processor, and a terminal pressure prediction control system having an input / output device and a memory device connected to the control unit, wherein the control unit: a) measures the supply side pressure using a single or multiple pressure gauges provided on the discharge side that supplies compressed air to the plurality of pipes; b) obtains pressure values ​​and flow values ​​when the valves are open or closed using pressure gauges, flow meters, and valves provided on all or some of the terminals of the pipes; c) learns a terminal pressure prediction model having a pressure for each terminal as an output using the supply pressure value from the air compressor and the pressure value and the flow value on the demand side; and during operation of the pneumatic system, the control unit: d) connects compressed air utilization equipment to the terminals in place of the valves; and e) calculates an optimal discharge pressure of the air compressor by inputting the supply pressure value and the flow value of the terminals into the terminal pressure prediction model.

10. The compressor control method according to claim 9, characterized in that one or more air compressors are provided, and the control unit adjusts the rotation speed and number of operating air compressors so that the air compressors can be operated at the optimal discharge pressure set in step e).

11. The compressor control method according to claim 10, characterized in that in step d), the control unit calculates future prediction values ​​of the multiple terminal pressures and the discharge pressure of the air compressor, and controls the rotation speed and number of operating air compressors so that the future prediction values ​​do not fall below a target value that is the lower limit of the terminal pressure.

12. The compressor control method described in claim 11, characterized in that the control unit uses many of the terminal pressure measurement values ​​and flow rate information when learning the terminal pressure prediction model in steps a) to c), and uses a smaller number of terminal flow rate information values ​​than the measurements in step a) when operating the compressed air utilization equipment in step e).

13. The compressor control method according to claim 9, further comprising providing a display device that visually displays information to a user, and displaying the optimal discharge pressure of the air compressor set by the control unit on the display device.

14. A compressor control method as described in claim 9, further comprising an input / output unit for inputting and outputting information, wherein the control unit simulates the optimal discharge pressure of the air compressor and displays it on the input / output unit in response to a user inputting the terminal target pressure from the input / output unit.

15. A compressor control method for a pneumatic system in which compressed air discharged from an air compressor is supplied through multiple pipes to multiple terminals that consume the compressed air, using a control unit having a processor, and a terminal pressure prediction control system having an input / output device and a storage device connected to the control unit, wherein multiple branching portions of the pipes are defined as zones, and a representative terminal that measures flow rate for each zone is set, wherein the control unit: a) measures a supply pressure value using a single or multiple pressure gauges provided on the discharge side that supplies compressed air to the multiple pipes, b) obtains pressure measurement values ​​and flow rate measurement values ​​when the valve is opened or closed using a pressure gauge, a flow meter, and a valve provided at the representative terminal, c) learns a terminal pressure prediction model having pressure for each terminal as an output using the supply pressure value and the pressure measurement value and the flow rate measurement value on the demand side, and during operation of the pneumatic system, the control unit: d) connects compressed air utilization equipment to the representative terminal instead of the valve, and e) calculates an optimal discharge pressure of the air compressor by inputting the supply pressure value and the flow rate measurement value for each representative terminal into the terminal pressure prediction model.

Images