Air compressor / refrigeration machine cluster control method and system

Abstract

The present invention discloses a cluster control method and system for air compressors / refrigerators, including: an air compressor / refrigerator cluster, a gas-using equipment / cold-using equipment cluster, a value acquisition module, an algorithm module, an accumulation module, an adjustment module, a control module, an allocation module, and a diagnosis module. The cluster control method and system for air compressors / refrigerators provided by the present invention obtains the start-up time and gas consumption / cold capacity of gas-using equipment / cold-using equipment by detecting a single gas-using equipment / cold-using equipment, detects power supply parameter information, accumulates and adjusts the operation conditions of all equipment after the periodic detection time, and obtains the gas consumption law of the equipment after the first startup by combining the back propagation error calculation and the self-learning function. The control module combines the fuzzy control and predictive control algorithms to form a control model, and gives control parameters according to the output data. The allocation module performs cluster-type dynamic adjustment, real-time feedback of operation information, and judgment of equipment failure and gas supply network leakage on the cluster system of gas-using equipment / cold-using equipment in the whole plant according to the above-obtained total gas consumption / total cold capacity.

Description

Technical Field

[0001] The present invention relates to the technical field of general utility engineering, and in particular to an air compressor / refrigeration machine cluster control method and system. Background Art

[0002] Compressed air is the most important industrial power source besides electricity, and is widely used in machinery, electronics, petrochemicals, automobiles, textiles, chemical fibers, non-ferrous metals, electricity, coal, food, medicine, etc. Air compressors (hereinafter referred to as air compressors) can provide compressed air as air source power. They are the core equipment of pneumatic systems and the main body of electromechanical air source devices. They are devices that convert the mechanical energy of the motive force into gas pressure energy and are compressed air pressure generating devices. At present, the compressed air system has the following problems: (1) The air volume adjustment range of centrifugal air compressors is small. Generally, a loading / unloading valve is used to control the air compressor to provide a large air volume to meet the gas consumption requirements of the production process. The frequent operation of high-pressure equipment not only has a large impact on the power grid, but also has a large loss on the life of components, and key components such as bearings are seriously worn; (2) There is a large pressure difference in the system. According to calculations, for every 0.1MPa increase in air pressure, the system will increase energy consumption by 7.5%, and the power consumption during standby is also 20%-50% of the load, which seriously wastes electricity; (3) The single-machine system cannot provide a fixed pressure according to the load conditions, or achieve energy-saving control through single-machine frequency conversion and standby; (4) The exhaust valve, operating frequency, etc. cannot be adjusted according to the system pipeline network conditions; (5) When a problem occurs, it is impossible to quickly find the cause of the fault and cannot timely and accurately feedback the equipment operation data, and it is impossible to judge faults such as pipeline leakage. Summary of the invention

[0003] In view of the deficiencies in the prior art, the present invention provides an air compressor / refrigeration machine cluster control method and system to perform cluster dynamic adjustment, real-time feedback of operation information, and judgment of equipment failure and gas supply network leakage.

[0004] The object of the present invention is achieved in the following manner: A method for controlling an air compressor / refrigeration machine cluster comprises the following steps:

[0005] Through the value acquisition module, a single gas-using device / cooling device is detected to obtain the opening time and gas / cooling amount of the gas-using device / cooling device, detect the power supply parameter information and perform error analysis;

[0006] The algorithm module can determine the gas / cooling and electricity characteristics of gas / cooling equipment in advance after the periodic detection time;

[0007] The operation status of all equipment is accumulated through the accumulation module to obtain the required total gas consumption / total cooling capacity;

[0008] Adjust the gas consumption / cooling capacity according to the pipe network conditions through the adjustment module, and increase the dead zone determination;

[0009] Control each air compressor / refrigeration machine cluster system through the control module, cluster dynamic adjustment and operation information feedback function;

[0010] According to the total gas consumption / total cooling capacity, the allocation module is used to dynamically allocate and wake up the type and quantity of the corresponding air compressor / refrigerator, and to perform steady-state control so that all running air compressors / refrigerators adopt the same frequency operation scheme, further increasing energy-saving efficiency and adding self-learning function;

[0011] The fault diagnosis module establishes corresponding mathematical models based on the algorithm module and the accumulation module to diagnose equipment failures and gas supply network leakage in real time;

[0012] The control module controls the cycle operation time of each device to increase the balanced life.

[0013] As an optional solution to the technical solution of the present invention, the value acquisition module detects the gas-using equipment / cold-using equipment cluster system through Profinet bus communication, obtains the equipment start-up time and gas / cold consumption, reads the pipeline pressure and flow through I / O, and reads the power supply side data through OPC.

[0014] As an optional solution of the technical solution of the present invention, the control module controls each air compressor / refrigeration machine cluster system through the Modbus RTU bus communication protocol, and at the same time performs remote cluster dynamic adjustment and operation information feedback functions of DCS through the Modbus TCP / IP industrial protocol system.

[0015] As an optional solution of the technical solution of the present invention, the total gas consumption / total cooling capacity accumulated by the accumulation module is calculated in the following way:

[0016] ;

[0017] in, is the total gas consumption / total cooling capacity, is the gas consumption / cooling capacity of a single device, The relationship between the compressor speed n, coefficient K and volume V is as follows: The adjustment module adjusts the coefficient according to the pipe network conditions. Adjustment of coefficient The adjustment adopts Asheim model, Yuan model, Ouyang Liang-Biao model or custom model according to the usage scenario, and then calculates the balance parameters and error analysis of each balance point.

[0018] As an optional solution of the technical solution of the present invention, it also includes: the specific calculation idea of ​​the integrated self-learning function of the allocation module is as follows:

[0019] (1) Normalization processing formula

[0020] ;

[0021] Where: It is the basic data after dimensionless transformation; is the SAGD well data corresponding to a certain factor; and The upper and lower limits of the data value;

[0022] (2) Forward calculation process

[0023] The components of the network input pattern are used as inputs to the nodes of the i-th layer (input layer). The outputs of the nodes in this layer are exactly equal to their input values, that is, ; The input value of the node in the jth layer (hidden layer) of the network is:

[0024] ;

[0025] The output value of the node in this layer is:

[0026] ;

[0027] in, is the output value of the j-th layer node, is the connection value between hidden layer node j and input layer node i, is the threshold of hidden layer node j, and f is the activation function of the node. Here we take the following monotonically increasing Sigmoid activation function:

[0028] ;

[0029] in, The role of is to change the shape of the activation function.

[0030] The input value of the kth layer (output layer) node of the network is , and the output value of the node in this layer is:

[0031] ;

[0032] in, is the output value of the k-th layer node, is the connection weight between the output layer node k and the hidden layer node j, is the threshold of the output layer node k, and f is the node activation function, which is generally taken as ;

[0033] (3) Error back propagation process

[0034] For the connection weight adjustment formula between the output layer and the hidden layer, there is the following formula:

[0035] ;

[0036] in, The error term of output layer node k, , The target output.

[0037] For the connection weight adjustment formula between the hidden layer and the input layer, there is the following formula:

[0038] ;

[0039] Right now

[0040] ;

[0041] in, is the connection weight between hidden layer node j and input layer node i, is the error term of hidden layer node j, , Indicates Iterations, is the inertia factor, which is used to adjust the convergence speed of network learning. The value will be helpful in suppressing the oscillation of the network. is the learning step size, that is, the weight gain factor.

[0042] As an optional solution of the technical solution of the present invention, the control module includes:

[0043] Start and stop control submodule: start and stop the corresponding equipment;

[0044] Frequency control submodule: Dynamic frequency modulation of variable frequency air compressor / refrigeration machine;

[0045] Guide vane opening control module: dynamically adjust the guide vane opening of the refrigerator;

[0046] Exhaust valve adjustment control module: dynamically adjust the opening of the air compressor / refrigeration machine exhaust valve.

[0047] In addition, the present invention also provides an air compressor / refrigeration machine cluster control system using the above control method.

[0048] An air compressor / refrigeration machine cluster control system, comprising: an air compressor / refrigeration machine cluster, an air-using equipment / refrigeration equipment cluster, a value acquisition module, an algorithm module, an accumulation module, an adjustment module, a control module, a distribution module and a fault diagnosis module;

[0049] The value acquisition module is connected with the gas-using equipment / cold-using equipment cluster and the algorithm module;

[0050] The algorithm module is connected with the accumulation module;

[0051] The accumulation module is connected with the adjustment module;

[0052] The adjustment module is connected to the control module;

[0053] The control module is connected to the distribution module;

[0054] The distribution module is connected to the air compressor / refrigeration machine cluster;

[0055] The fault diagnosis module is connected with the control module and the accumulation module;.

[0056] As an optional solution of the technical solution of the present invention, the algorithm module includes:

[0057] Computation submodule for periodic adaptation;

[0058] The change amount module is used to adapt the change amount of gas consumption / cooling capacity at each stage of an operation cycle;

[0059] The power data submodule is used to read power supply parameters and analyze power consumption history data, and provide it to the calculation submodule.

[0060] As an optional solution of the technical solution of the present invention, the control module includes:

[0061] The start-stop control module starts and stops the corresponding equipment;

[0062] Frequency control submodule: Dynamically adjust the frequency of variable frequency air compressor / refrigeration machine clusters;

[0063] Guide vane opening control module: dynamically adjusts the guide vane opening of the air compressor / refrigeration machine cluster;

[0064] Exhaust valve adjustment control module: dynamically adjust the exhaust valve opening of the air compressor / refrigeration machine cluster;

[0065] The fault diagnosis module comprises:

[0066] Fault analysis module: performs fault diagnosis on the air compressor / refrigeration machine cluster based on the data reading module;

[0067] Leakage diagnosis module: establish corresponding mathematical models for the above algorithm module and accumulation module to perform real-time diagnosis of leakage.

[0068] The beneficial effects of the present invention are:

[0069] The device for cluster control of air compressors / refrigeration machines provided by the present invention has the beneficial effect that it is installed in a control room and connected to air compressors and field equipment. The value acquisition module communicates with the equipment end through Profinet bus communication, reads power supply side data through OPC, and reads pipe network pressure and flow through I / O. After calculation, accumulation and adjustment, the control module controls each air compressor / refrigeration machine cluster system through the Modbus RTU bus communication protocol. At the same time, through the DCS remote parameter adjustment and operation information feedback functions of industrial protocol systems such as Modbus TCP / IP, the cluster system of gas-using equipment / refrigeration equipment in the whole plant performs cluster dynamic adjustment, real-time feedback of operation information, judgment of equipment failure and gas supply network leakage, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0071] Figure 1 This is a flow chart of a method for controlling an air compressor cluster in Embodiment 1 of the present invention;

[0072] Figure 2 This is a logic diagram of the air compressor cluster control method in Embodiment 1 of the present invention;

[0073] Figure 3 This is a hardware schematic diagram of the air compressor cluster control system in Embodiment 1 of the present invention;

[0074] Figure 4 This is a schematic diagram of a value acquisition module in an air compressor control method and system in Embodiment 1 of the present invention;

[0075] Figure 5 This is a schematic diagram of a control module in an air compressor cluster control method and system in Embodiment 1 of the present invention;

[0076] Figure 6 Schematic diagram of the air compressor cluster control system in Example 1 of the present invention. DETAILED DESCRIPTION

[0077] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0078] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0079] In the description of the embodiments, unless otherwise clearly specified and limited, the terms "disposed", "connected" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, or an electrical connection; it can be a direct connection, or it can be connected through an intermediate medium, or it can be a communication between the two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0080] Embodiment 1

[0081] like Figure 1-5 As shown, this embodiment discloses a method for controlling an air compressor cluster, comprising the following steps:

[0082] Value acquisition module: Through the detection of a single device, the start time and gas consumption of the gas-consuming equipment are obtained, the power supply parameter information is detected and error analysis is performed;

[0083] Algorithm module: After the periodic detection time, the gas consumption pattern of the device after the first startup is obtained by combining the back propagation error calculation; at a startup time T, the power consumption EPI, the gas consumption Q, the specific energy α=(EPI*T) / Q, combined with the self-learning function, the convergence speed of the detection time is reduced. Since the device EPI and T are difficult to adjust, α can be reduced by adjusting Q;

[0084] Accumulation module: accumulates the operation status of all equipment to obtain the required total gas consumption; the total gas consumption accumulated by the accumulation module is calculated in the following way:

[0085] ;

[0086] in, is the total gas consumption, is the gas consumption of a single device, The relationship between the compressor speed n, coefficient K and volume V is as follows: ;

[0087] Adjustment module: Adjust the gas consumption according to the pipeline network conditions, and add dead zone determination; the adjustment module adjusts the coefficient according to the pipeline network conditions, and adjusts the coefficient according to the usage scenario by

[0088] Asheim model: , Yuan model: , Ouyang Liang-Biao: and other three models, and custom models can also be defined. Then calculate the balance parameters: , and the error analysis of each balance point: ;

[0089] Control module: Combine the algorithms of fuzzy control and predictive control to form a control model, and give control parameters according to the output data. Because is the - output by the air compressor, and actually need to control the air compressor output >> , the control module can adjust the rotational speed n of the air compressor to adjust ;

[0090] Distribution module: According to the total gas consumption required, dynamically allocate through the distribution module to wake up the types and quantities of the corresponding air compressors to run, and perform steady-state control, so that all running air compressors adopt the same frequency operation scheme, further increasing the energy-saving efficiency, and adding a self-learning function.

[0091] The specific calculation idea is as follows:

[0092] Let X1 be the current number of started units, X2 be the allocated number of started units, and X1 - X2 represents the difference between the current number of started units and the allocated number of started units. Set the threshold m according to the actual operation of the air compressor. During actual allocation, if 0 < |X1| - |X2| < m, then the rotational speed n of the air compressor needs to be gradually adjusted. If m < |X1| - |X2|, then the air compressor needs to be gradually shut down.

[0093] The specific calculation idea of the self-learning function integrated in the distribution module is as follows:

[0094] (1) Normalization processing formula

[0095] ;

[0096] In the formula: is the dimensionless basic data; is the SAGD well data corresponding to a certain factor; and are the upper and lower limits of the data;

[0097] (2) Forward calculation process

[0098] The components of the network input pattern are used as inputs to the nodes of the i-th layer (input layer). The outputs of the nodes in this layer are exactly equal to their input values, that is, ; The input value of the node in the jth layer (hidden layer) of the network is:

[0099] ;

[0100] The output value of the node in this layer is:

[0101] ;

[0102] in, is the output value of the j-th layer node, is the connection value between hidden layer node j and input layer node i, is the threshold of hidden layer node j, and f is the activation function of the node. Here we take the following monotonically increasing Sigmoid activation function:

[0103] ;

[0104] in, The role of is to change the shape of the activation function;

[0105] The input value of the kth layer (output layer) node of the network is , and the output value of the node in this layer is:

[0106] ;

[0107] in, is the output value of the k-th layer node, is the connection weight between the output layer node k and the hidden layer node j, The threshold of the output layer node k, f is the node activation function, generally taken as ;

[0108] (3) Error back propagation process

[0109] For the connection weight adjustment formula between the output layer and the hidden layer, there is the following formula:

[0110] ;

[0111] in, is the error term of output layer node k, , The target output.

[0112] For the connection weight adjustment formula between the hidden layer and the input layer, there is the following formula:

[0113] ;

[0114] Right now

[0115] ;

[0116] in, is the connection weight between hidden layer node j and input layer node i, is the error term of hidden layer node j, .here Indicates Iterations, is the inertia factor, which is used to adjust the convergence speed of network learning. The value will be helpful in suppressing network oscillation. is the learning step size, that is, the weight gain factor.

[0117] Fault diagnosis module: Establish corresponding mathematical models based on the algorithm module and accumulation module, and timely judge the fault point of the air compressor based on the motor parameters, oil pump parameters, temperature, pressure, surge and other parameters of the air compressor; make timely diagnosis of pipe network leakage based on flow information;

[0118] The control module accumulates the running time of each device and increases the balanced life.

[0119] Specifically,

[0120] The control module includes:

[0121] Start and stop control submodule: start and stop the corresponding equipment;

[0122] Frequency control submodule: dynamic frequency regulation of variable frequency air compressor;

[0123] Exhaust valve adjustment control module: dynamically adjust the opening of the air compressor exhaust valve.

[0124] like Figure 3 As shown, the value acquisition module detects the gas-consuming equipment cluster system through Profinet bus communication, obtains the equipment start-up time and gas consumption, reads the pipeline pressure and flow through I / O, and reads the power supply side data through OPC. After calculation, accumulation and adjustment, the control module controls each air compressor cluster system through the Modbus RTU bus communication protocol, and at the same time performs remote cluster dynamic adjustment and operation information feedback functions through the Modbus TCP / IP industrial protocol system for DCS.

[0125] In addition, if Figure 6As shown, the present invention also provides an air compressor cluster control device using the above control method, comprising: an air compressor cluster, a gas-using equipment cluster, a value acquisition module, an algorithm module, an accumulation module, an adjustment module, a control module and a distribution module, wherein:

[0126] The value acquisition module is connected with the gas consumption equipment cluster and the algorithm module;

[0127] Value acquisition module: By detecting a single gas-consuming device, the start-up time and gas consumption of the gas-consuming device are obtained, and the power supply parameter information is detected; specifically, the start-up time and gas consumption of the gas-consuming device are obtained through Profinet bus communication, the pipeline pressure and flow are read through I / O, and the power supply side data is read through OPC.

[0128] The algorithm module is connected with the accumulation module;

[0129] Algorithm module: After periodic detection time, the gas consumption pattern of the gas-consuming equipment after the first startup is obtained by combining the back propagation error calculation;

[0130] The accumulation module is connected with the adjustment module;

[0131] Accumulation module: accumulates the operation status of the gas-consuming equipment cluster;

[0132] The adjustment module is connected with the control module;

[0133] Adjustment module: The coefficient K can be adjusted according to the pipe network conditions, where: The following relationship exists between the compressor speed n, coefficient K and volume V: =K*V*n;

[0134] The control module is connected to the distribution module;

[0135] Control module: Combine fuzzy control and predictive control algorithms into a control model, give control parameters based on output data, and adjust the air compressor speed n ;

[0136] The distribution module is connected to the air compressor cluster;

[0137] Allocation module: According to the required total gas consumption obtained above, dynamic allocation is performed to wake up the corresponding air compressor operation type and quantity.

[0138] The fault diagnosis module is connected with the control module and the accumulation module.

[0139] Fault diagnosis module: timely determine the fault point of the air compressor based on the motor parameters, oil pump parameters, temperature, pressure, surge and other parameters of the air compressor; timely diagnose the leakage of the pipeline network based on the flow information; and timely diagnose the leakage of gas-using equipment based on the historical information of gas consumption of a single device.

[0140] Specifically, the algorithm modules include:

[0141] Computation submodule for periodic adaptation;

[0142] The change amount module is used to adapt the change amount of gas consumption at each stage in an operation cycle;

[0143] The power data submodule is used to read power supply parameters and analyze power consumption history data, and provide it to the calculation submodule.

[0144] The control module includes:

[0145] The start-stop control module starts and stops the corresponding equipment;

[0146] Frequency control submodule: dynamically adjust the frequency of variable frequency air compressor clusters;

[0147] Guide vane opening control module: dynamically adjusts the guide vane opening of the air compressor cluster;

[0148] Exhaust valve adjustment control module: dynamically adjusts the exhaust valve opening of the air compressor cluster.

[0149] The fault diagnosis module includes:

[0150] Fault analysis module: performs fault diagnosis on the air compressor / refrigeration machine cluster based on the data reading module;

[0151] Leakage diagnosis module: establish corresponding mathematical models for the above algorithm module and accumulation module to perform real-time diagnosis of leakage.

[0152] In addition, the control module is also connected to an HMI touch panel to facilitate the adjustment of various parameters.

[0153] This control method is applicable to the control of an air compressor cluster, which may be a screw air compressor, a centrifugal air compressor or a cluster of multiple air compressors.

[0154] Embodiment 2

[0155] Of course, the control method and system of the present invention are also applicable to the control of a refrigeration machine cluster. The difference between this embodiment and embodiment one is that the refrigeration machine cluster control module includes a guide vane opening control module for dynamically adjusting the guide vane opening of the refrigeration machine, the air compressor cluster control method controls the air volume of the air compressor, and the refrigeration machine cluster control method controls the cooling capacity of the refrigeration machine.

[0156] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement; when the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the protection scope required by the present invention.

Claims

1. A method for controlling an air compressor / refrigeration machine cluster, characterized in that: The following steps are involved: Through the value acquisition module, a single gas-using device / cooling device is detected to obtain the opening time and gas / cooling amount of the gas-using device / cooling device, detect the power supply parameter information and perform error analysis; The algorithm module can determine the gas / cooling and electricity characteristics of gas / cooling equipment in advance after the periodic detection time; The operation status of all equipment is accumulated through the accumulation module to obtain the required total gas consumption / total cooling capacity; Adjust the gas consumption / cooling capacity according to the pipe network conditions through the adjustment module, and increase the dead zone determination; Control each air compressor / refrigeration machine cluster system through the control module, cluster dynamic adjustment and operation information feedback function; According to the required total gas consumption / total cooling capacity, the allocation module is used to dynamically allocate and wake up the type and quantity of the corresponding air compressor / refrigerator, and to perform steady-state control so that all running air compressors / refrigerators adopt the same frequency operation scheme, further increasing energy-saving efficiency and adding self-learning function; The fault diagnosis module establishes corresponding mathematical models based on the algorithm module and the accumulation module to diagnose equipment failures and gas supply network leakage in real time; The control module controls the cycle operation time of each device to increase the balanced life.

2. The air compressor / refrigeration machine cluster control method according to claim 1, characterized in that: The value acquisition module detects the gas-using equipment / cold-using equipment cluster system through Profinet bus communication, obtains the equipment start-up time and gas / cold consumption, reads the pipe network pressure and flow through I / O, and reads the power supply side data through OPC.

3. The air compressor / refrigeration machine cluster control method according to claim 1, characterized in that: The control module controls each air compressor / refrigeration machine cluster system through the Modbus RTU bus communication protocol, and at the same time performs remote cluster dynamic adjustment and operation information feedback functions of the DCS through the Modbus TCP / IP industrial protocol system.

4. The air compressor / refrigeration machine cluster control method according to claim 1, characterized in that: The total gas consumption / total cooling capacity accumulated by the accumulation module is calculated in the following way: ； in, is the total gas consumption / total cooling capacity, is the gas consumption / cooling capacity of a single device, The following relationship exists between the compressor speed n, coefficient K and volume V The adjustment module adjusts the coefficient according to the pipe network conditions. Adjustment of coefficient The adjustment adopts Asheim model, Yuan model, Ouyang Liang-Biao model or custom model according to the usage scenario, and then calculates the balance parameters and error analysis of each balance point.

5. According to the air compressor / refrigeration machine cluster control method of claim 1, the specific calculation idea of ​​the allocation module integrated self-learning function is as follows: (1) Normalization processing formula ； Where: It is the basic data after dimensionless transformation; is the SAGD well data corresponding to a certain factor; and The upper and lower limits of the data value; (2) Forward calculation process The components of the network input pattern are used as inputs to the nodes in the i-th layer, and the outputs of the nodes in this layer are completely equal to their input values, that is, ;The input value of the node in the jth layer of the network is: ； The output value of the node in this layer is: ； in, is the output value of the j-th layer node, is the connection value between hidden layer node j and input layer node i, is the threshold of hidden layer node j, and f is the activation function of the node. Here we take the following monotonically increasing Sigmoid activation function; ； in, The role of is to change the shape of the activation function; The input value of the k-th layer node of the network is , and the output value of the node in this layer is: ； in, is the output value of the k-th layer node, is the connection weight between the output layer node k and the hidden layer node j, is the threshold of the output layer node k, f is the node activation function, and ; (3) Error back propagation process For the connection weight adjustment formula between the output layer and the hidden layer, there is the following formula: ； in, is the error term of output layer node k, ; Output for the target; For the connection weight adjustment formula between the hidden layer and the input layer, there is the following formula: ； Right now ； in, is the connection weight between hidden layer node j and input layer node i, is the error term of hidden layer node j, , Indicates Iterations, is the inertia factor, which is used to adjust the convergence speed of network learning. The value will be helpful in suppressing the oscillation of the network. is the learning step size, that is, the weight gain factor.

6. The air compressor / refrigeration machine cluster control method according to claim 1, characterized in that: The control module comprises: Start and stop control submodule: start and stop the corresponding equipment; Frequency control submodule: Dynamic frequency modulation of variable frequency air compressor / refrigeration machine; Guide vane opening control module: dynamically adjust the guide vane opening of the refrigerator; Exhaust valve adjustment control module: dynamically adjust the opening of the air compressor / refrigeration machine exhaust valve.

7. An air compressor / refrigeration machine cluster control system, characterized in that: A method for controlling an air compressor / refrigeration machine cluster using any one of claims 1-6.

8. The air compressor / refrigeration machine cluster control system according to claim 7, characterized in that: include: Air compressor / refrigeration machine cluster, gas equipment / refrigeration equipment cluster, value acquisition module, algorithm module, accumulation module, adjustment module, control module, allocation module and fault diagnosis module; The value acquisition module is connected with the gas-using equipment / cooling equipment cluster and the algorithm module; The algorithm module is connected with the accumulation module; The accumulation module is connected with the adjustment module; The adjustment module is connected to the control module; The control module is connected to the distribution module; The fault diagnosis module is connected with the control module and the accumulation module.

9. The air compressor / refrigeration machine cluster control system according to claim 8, characterized in that: The algorithm module includes: Computation submodule for periodic adaptation; The change amount module is used to adapt the change amount of gas consumption / cooling capacity at each stage of an operation cycle; The power data submodule is used to read power supply parameters and analyze power consumption history data, and provide it to the calculation submodule.

10. The air compressor / refrigeration machine cluster control system according to claim 9, characterized in that: The control module comprises: The start-stop control submodule starts and stops the corresponding equipment; Frequency control submodule: Dynamically adjust the frequency of variable frequency air compressor / refrigeration machine clusters; Guide vane opening control module: dynamically adjusts the guide vane opening of the air compressor / refrigeration machine cluster; Exhaust valve adjustment control module: dynamically adjust the exhaust valve opening of the air compressor / refrigeration machine cluster; The fault diagnosis module comprises: Fault analysis module: performs fault diagnosis on the air compressor / refrigeration machine cluster based on the data reading module; Leakage diagnosis module: establish corresponding mathematical models for the above algorithm module and accumulation module to perform real-time diagnosis of leakage.

Images