A pipeline lean production management system and method thereof

Abstract

The application discloses a pipeline lean production management system and method, and relates to the technical field of ship manufacturing. The system solves the problems existing in the traditional production mode through workshop reconstruction and process optimization. The system adopts a hierarchical real-time control architecture, which comprises an edge computing gateway, a MES server and a data acquisition module. The edge computing gateway runs a hard real-time operating system and executes millisecond-level control logic. The MES server is arranged on a management layer, runs a general operating system, and executes second-level monitoring and historical analysis. The data acquisition module is arranged on production equipment of a core machining process of a pipeline workshop and is connected with the edge computing gateway and the MES server.

Description

Technical Field

[0001] This invention relates to the field of shipbuilding technology, specifically to a pipeline lean production management system and method. Background Technology

[0002] In the shipbuilding industry, pipe prefabrication is a core production step connecting critical equipment. Traditional pipework workshops generally employ a multi-functional, hybrid production model, which presents the following prominent problems:

[0003] Low production efficiency: The mixed processing of pipes made of different materials (stainless steel, carbon steel, copper pipe) leads to frequent switching of production tasks. The time for replacing welding shielding gas and purging pipes alone is 20-40 minutes per time, and the overall equipment utilization rate (OEE) is generally less than 60%.

[0004] Serious waste of resources: In the mixed production mode, tooling changeover time is as long as 28 minutes / time (e.g., the weight difference between Φ216mm and Φ60.5mm bending molds is 80kg, and manual changeover is time-consuming). There is serious waste of work-in-process waiting and handling between processes, and the inventory turnover rate is less than 50% of the industry average.

[0005] Outdated production organization methods: The sequential movement method, which is suitable for mass production, is seriously mismatched with the multi-variety, small-batch production needs of ship conversion projects, and the production cycle cannot be accurately predicted.

[0006] Bottleneck process identification is delayed: Traditional manual experience is used to identify bottleneck processes, with a response delay of 2-4 hours, and it is impossible to capture in real time the process cycle drift caused by welding machine aging and tooling wear.

[0007] While existing technologies mention MES systems and lean manufacturing, they have not solved the following technical problems:

[0008] There is a lack of physical isolation and reconstruction solutions tailored to the characteristics of pipeline materials;

[0009] Bottleneck process identification relies solely on a single equipment utilization rate indicator without considering dynamic changes in the work-in-process queue, resulting in a high false alarm rate.

[0010] The choice of production organization methods relies on manual experience and does not form a closed loop with real-time order data and production line configuration.

[0011] Therefore, there is an urgent need in this field for a technical system that deeply integrates workshop physical reconstruction, hard real-time data acquisition, dynamic bottleneck process and bottleneck capacity production line algorithms, and intelligent decision-making logic, in order to achieve a fundamental shift from "experience-driven" to "data-driven". Summary of the Invention

[0012] One of the objectives of this invention is to propose a lean production management method for pipelines, which solves the problems existing in the traditional production model through workshop restructuring and process optimization.

[0013] The technical solution of the present invention is as follows:

[0014] A pipeline lean production management system and method, employing a hierarchical real-time control architecture, includes:

[0015] Edge computing gateways run hard real-time operating systems and execute millisecond-level control logic.

[0016] The MES server is deployed at the management level, runs a general operating system, and performs second-level monitoring and historical analysis.

[0017] The data acquisition module is deployed on the production equipment of the core processing steps in the pipeline workshop, and connects the edge computing gateway and the MES server;

[0018] The edge computing gateway communicates with the MES server and collaboratively executes the following steps:

[0019] S1: Physically restructure the pipeline workshop from a mixed production area for multiple materials into a dedicated work area for a single material, and divide it into m types of flexible production lines {A1, A2, ..., A...} according to pipe diameter specifications. m The number of production lines of type i, Aᵢ, is Bᵢ.

[0020] S2: Based on the real-time data from the data acquisition module, bottleneck processes are dynamically identified through a bottleneck determination algorithm. The algorithm is as follows: when the equipment utilization rate of a certain process is >80% and the number of work-in-process at its entry point continues to increase, the process is determined to be a bottleneck.

[0021] S3: Based on the bottleneck identification results and the total installation requirement R, calculate the theoretical production batch nᵢ=Rᵢ / (single line capacity×Bᵢ), and automatically select the production organization method: if nᵢ≤1, adopt small batch one-pass flow operation; if nᵢ>1, adopt parallel sequential moving operation.

[0022] S4: Generates control commands and sends them to the production line PLC via the edge computing gateway to execute production.

[0023] Furthermore, the edge computing gateway also performs cyclic batch production control:

[0024] When the theoretical production batch size nᵢ>x (x is the preset maximum batch size, x≥3), a cyclic batch production mode is adopted, with each batch having a quantity of x and a production cycle of T. x ;

[0025] Total production cycle Tz = (n / x)·T xThe edge computing gateway automatically triggers the next batch production instruction after each batch is completed, without the need for manual intervention.

[0026] Furthermore, the data acquisition module is configured hierarchically according to process type:

[0027] Core process data acquisition unit: Sensors and timers installed on the blanking machine for detecting position, acquiring cutting length and processing time t1; sensors installed on the pipe bending machine for detecting angle and pressure, acquiring bending angle, forming pressure and processing time t2; sensors installed on the welding machine for detecting current and gas flow rate, acquiring welding parameters and processing time t4.

[0028] Auxiliary process data acquisition unit: RFID readers are deployed in the assembly process, visual inspection process, film shooting process, and assembly delivery process to record the start / end time t3, t5, t6, and t7 of the process by scanning the work-in-process pallet tags in order to monitor the increase or decrease of the quantity of work-in-process;

[0029] The workshop industrial Ethernet transmits the aforementioned sensor and RFID data to the edge computing gateway and MES server in real time.

[0030] Furthermore, in the hierarchical real-time control architecture:

[0031] The edge computing gateway maintains a lightweight state mapping table (memory variable) for millisecond-level bottleneck determination;

[0032] The MES server stores a complete real-time state mapping model (SQL database) for second-level monitoring and historical backtracking.

[0033] The real-time state mapping model includes: a real-time queue Qᵢ and a bottleneck marker register;

[0034] The edge gateway and the MES server maintain model consistency through an asynchronous data synchronization protocol (such as MQTT).

[0035] Furthermore, the bottleneck determination algorithm executed by the edge computing gateway is specifically as follows:

[0036] The operating status of each process equipment is read from the data acquisition module according to the configurable sampling period Δt;

[0037] The equipment utilization rate Uᵢ for process Mᵢ is calculated as: actual processing time / total observation time.

[0038] When Uᵢ>80% and Qᵢ(t+Δt)>Qᵢ(t) for three consecutive sampling periods, Mᵢ is marked as a bottleneck process. Specifically, the edge computing gateway sets the bottleneck marker register of Mᵢ and directly triggers the PLC physical address through the hard real-time industrial communication protocol. After the MES server synchronizes the marker via asynchronous message, it triggers the Kanban audible and visual alarm and pushes the work order to the production supervisor's mobile terminal.

[0039] Furthermore, in step S3 executed by the edge computing gateway, the bottleneck capacity identification and production batch determination method is as follows:

[0040] The total installation requirements R given by the shipboard installation team are resolved into a set of corresponding requirements for various types of piping {R1, R2, ..., R...}. m};

[0041] Based on the number of production lines Aᵢ of type i, Bᵢ and their single-line capacity, calculate in real time the theoretical production batches nᵢ required to meet the various pipeline requirements Rᵢ, which is calculated as nᵢ=Rᵢ / (single-line capacity×Bᵢ).

[0042] The theoretical production batch n of the system is determined to be the maximum value among all nᵢ, that is, n=max(nᵢ);

[0043] The production line corresponding to the maximum value nᵢ is dynamically marked as the bottleneck capacity production line. Specifically, the bottleneck mark field of the production line is set in the SQL database of the MES server, and the production cycle of the bottleneck line is used as the production cycle benchmark for the pipeline workshop's delivery order requirement R.

[0044] Furthermore, the logic for automatically selecting the production organization method in step S3 executed by the edge computing gateway is as follows:

[0045] The theoretical production batch size of the calculation system is n = max(nᵢ).

[0046] If n≤1, the edge computing gateway generates a single streaming job instruction, with a production cycle T. op This is the sum of the cycle times for each process.

[0047] If n>1, the edge computing gateway generates parallel sequential movement instructions and dynamically calculates the optimal production cycle T based on the process cycle time. op .

[0048] Furthermore, the parallel sequential movement instructions generated by the edge computing gateway include timing offsets for inter-process transfers:

[0049] When the inter-process single-piece processing time tᵢ≥ti+1 is detected, the instruction embedding offset Δtᵢ=(x-1)·ti+1;

[0050] The offset is used to precisely control the start time of the batch in the next process, avoiding idle waiting of the equipment.

[0051] Another object of the present invention is to provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to perform the functions of the system as described in any one of claims 1 to 8.

[0052] Another object of the present invention is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the functions of the system as described in any one of claims 1 to 8.

[0053] The present invention achieves the above objectives through the following technical means:

[0054] 1. Hierarchical real-time control architecture: The edge computing gateway performs millisecond-level hard real-time control, while the MES server performs second-level soft real-time monitoring. The two work together through an asynchronous data synchronization protocol.

[0055] 2. Physical Reconstruction and Data Acquisition: The mixed production area is reconstructed into a dedicated work area for a single material. Flexible production lines are divided into m categories according to pipe diameter specifications. Sensors are deployed in core processes (material cutting, pipe bending, welding), and RFID (data readers) are deployed in auxiliary processes (assembly, visual inspection, film taking, assembly) to build a hard real-time data acquisition network.

[0056] 3. Value Stream Analysis and Waste Elimination: By drawing the current state value stream map, the system identifies 7 typical wastes (waiting, handling, motion, defective products, overproduction, overprocessing, and inventory) in the entire process from material supply to shipboard installation. It designs the future state value stream map and eliminates non-value-adding links by establishing a cyclical pull production mechanism and a timed delivery system, effectively eliminating process bottlenecks and waiting time.

[0057] 4. Hard real-time bottleneck determination: The edge gateway is based on the dual conditions of device utilization >80% and continuous queue growth. After verification through 3 consecutive sampling cycles (90 seconds), the false alarm rate is reduced to 2.1%, and the PLC physical address is directly triggered with a response delay of <1ms.

[0058] 5. Selection of intelligent production organization method: The edge gateway is based on the theoretical production batch... Automatically select one-time flow job ( ) or parallel sequential movement ( ), and introduced a cyclic batch mode ( The batch transition is automatically triggered by the state machine, forming a closed loop.

[0059] 6. Timing Offset Control: In parallel sequential movement, an offset is embedded when tᵢ≥ti+1. Precisely control the start time of the process to avoid equipment idling and waiting.

[0060] 7. Standard production organization formula:

[0061] Sequential movement: Production cycle (The sum of the cycle times of each process);

[0062] Parallel sequential movement:

[0063] The production cycle formula is:

[0064] ;

[0065] In the formula: For batches; Number of processes; For the first The processing time per piece in a process.

[0066] Compared with the prior art, the present invention has the following significant advantages:

[0067] 1) Production cycle time was reduced by 40.7% (from 630 minutes to 373.3 minutes);

[0068] 2) The accuracy of bottleneck process identification has been improved to 100% (compared to 60% in the traditional method);

[0069] 3) Equipment utilization rate increased to 79% (an increase of 36%), and work-in-process inventory decreased to 4.5 units (a decrease of 73%).

[0070] 4) Mold change time reduced to 9 minutes (80% reduction);

[0071] 5) On-time delivery rate improved to 96.5% (an increase of 34%). Attached Figure Description

[0072] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0073] Figure 1 This is the overall architecture diagram of the pipeline lean production management system of this invention;

[0074] Figure 2 This is a top view of the physical reconstruction layout of the workshop according to the present invention;

[0075] Figure 3This is a schematic diagram of the real-time state mapping model structure of the present invention, used to illustrate the hierarchical mapping relationship between physical production line nodes and MES virtual nodes. The upper table shows the process data flow, and the lower column shows the time characteristic comparison. The hard real-time domain runs in the memory of the edge gateway and is synchronized to the SQL database of the MES server every 30 seconds via the MQTT protocol, forming the soft real-time domain. When the welding process triggers a bottleneck, the edge gateway directly writes to the PLC address DB100.DBX0.0, with a response latency of less than 1 millisecond.

[0076] Figure 4 This is a flowchart of the bottleneck determination algorithm of the present invention;

[0077] Figure 5 This is a timing diagram of production batch calculation and bottleneck identification in this invention;

[0078] Figure 6 This is a schematic diagram of the parallel sequential movement timing offset of the present invention.

[0079] Appendix Number Attachment title Corresponding claims Core content and technical features Figure 1 Overall architecture diagram of the pipeline lean production management system; hardware implementation of processor execution logic. Claim 1 The presentation showcases a three-layer architecture consisting of a data acquisition layer, an MES server layer, and a processor layer, clearly defining the seven processes M1~M7, the data / control flow direction, and the hardware entities (sensors, RFID, PLC, and dashboards). Figure 2 Top view of the physical reconstruction layout of the workshop Claim 3 <![CDATA[Use colors to distinguish three single-material working areas, and mark the positions, spacings, and tracks of movable devices of flexible production lines A1 to A m . Mark sensors with ☆ and RFID with □]]> Figure 3 Real-time state mapping model structure diagram Claim 4 The diagram illustrates the correspondence between physical production line nodes and MES virtual nodes, presenting the bottleneck marker register and PLC address mapping in tabular form. Figure 4 Bottleneck determination algorithm flowchart Claims 5 and 6 Diamond-shaped decision box: Uᵢ>80%→Qᵢ(t+Δt)>Qᵢ(t)→Counter≥3, Output branch: Mark bottleneck→Trigger alarm / Push work order Figure 5 Production batch calculation and bottleneck identification; cyclic production control status Claims 2 and 6 The horizontal axis represents time, and the vertical axis is divided into three columns: order demand R, theoretical batch size nᵢ, and bottleneck marker. This displays the calculation process of max(nᵢ) and highlights the bottleneck production line. The UML state machine is as follows: Idle → Batch Calculation → Loop Execution (n / x) times → End, with automatic triggering indicated upon completion of each batch. Figure 6 Schematic diagram of parallel sequential movement and its offset Claims 7 and 8 <![CDATA[Graphical display of the transfer of batches between processes, marking the offset Δtᵢ=(x - 1)·t i+1 Calculation and delay effect]]>

[0080] In the diagram: 100, Pipeline workshop; 201, Sensor; 202, RFID; 210, Mobile equipment track; 211, Safety barrier;

[0081] M1. Material cutting process; M2. Pipe bending process; M3. Assembly process; M4. Welding process; M5. Visual inspection process; M6. X-ray inspection process; M7. Assembly and delivery process. Detailed Implementation

[0082] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0083] Example 1: System Hardware Deployment and Hierarchical Architecture Implementation

[0084] This embodiment details the hardware deployment scheme of the hierarchical real-time control architecture of claim 1.

[0085] 1. The data acquisition module is configured within the shipbuilding stainless steel piping workshop 100. The original three mixed production lines were dismantled and moved to the new area along pre-embedded steel plate tracks, reconstructing a dedicated work area for a single material. The specific deployment of sensor 201 is as follows:

[0086] Core process data acquisition unit:

[0087] In the material feeding process M1, the cutting head is equipped with a laser displacement sensor 201 (model: Keyence IL-300, accuracy ±0.1mm) and a PLC timer to collect the cutting length L and processing time t1.

[0088] An angle encoder (model: Omron E6B2-CWZ6C, resolution 3600P / R) and a sensor 201 (model: IFM PN3071, range 0-400bar) for detecting hydraulic pressure are installed on the bending die base of the pipe bending process M2 to collect the bending angle θ, forming pressure P and processing time t2.

[0089] The welding torch clamping mechanism of welding process M4 is equipped with a Hall sensor 201 (model: LEM IT400-S, range 0-500A) for detecting current and a gas flow meter (model: Qixing Huachuang CS200, range 0-30L / min) to collect welding parameters and processing time t4.

[0090] Auxiliary process data acquisition unit:

[0091] RFID202 readers (model: Siemens RF260R) are deployed in assembly process M3, visual inspection process M5, film inspection process M6, and assembly delivery process M7 to record the start / end times t3, t5, t6, and t7 of the process by scanning the work-in-process pallet labels.

[0092] 2. Network and Server Configuration

[0093] An industrial Ethernet switch (model: Hirschmann MS20-1600) is deployed. Data from sensor 201 and RFID 202 is aggregated via the industrial Ethernet protocol (Modbus TCP in this example; those skilled in the art can also use EtherNet / IP or Profinet) to an edge computing gateway (Advantech WISE-Edge-342, running VxWorks 7.0 real-time operating system) and an MES server (Dell PowerEdge R740, Windows Server 2019). A lightweight state mapping table is built in the gateway's memory, and a complete database model is built in the server. The two are synchronized asynchronously (using the MQTT protocol in this example) to maintain consistency.

[0094] 3. Data Flow in a Hierarchical Architecture

[0095] Downlink control flow: Gateway → PLC (hard real-time industrial communication protocol, response delay <1ms);

[0096] Uplink monitoring stream: Gateway → MES (asynchronous message synchronization, period 30 seconds);

[0097] Independent operation capability: The gateway has built-in power failure protection memory, so even if the MES server crashes, it can still independently perform bottleneck judgment and device control.

[0098] Example 2: Physical Reconstruction and Flexible Configuration of Stainless Steel Piping Workshop

[0099] This embodiment details the specific implementation process of step 1S1 of claim 1 for the stainless steel pipeline workshop 100.

[0100] 1. Historical order data analysis: 327 VLGC-type stainless steel piping orders from our factory between 2020 and 2022 were retrieved. Statistical analysis revealed:

[0101] The processing time ratio for large-diameter (≥Φ216mm), medium-diameter (Φ73~Φ168mm), and small-diameter (≤Φ60.5mm) pipes remained stable at 1:2.3:2.8, with a pipe diameter specification dispersion of 80%.

[0102] 2. Production line classification and quantity determination

[0103] Based on data analysis, the flexible production line for stainless steel pipelines is divided into three categories (such as...). Figure 2 (Layout shown)

[0104] Category A1: Large-diameter production line (pipe diameter ≥ Φ216mm), configured with B1=1 line, single line daily capacity of 9 pipes, located in the north area of ​​the workshop, 12 meters away from the east wall, with a reserved logistics channel width of 4 meters;

[0105] A2 type: medium diameter production line (Φ73~Φ168mm), configured with B2=2 lines, single line daily capacity of 9 pieces, double line daily capacity of 18 pieces, the two lines are laid out in parallel, the center distance is 8 meters, and a common tooling buffer area is set in the middle.

[0106] Category A3: Small-diameter production line (pipe diameter ≤ Φ60.5mm), configured with B3=3 lines, single line daily capacity of 9 pieces, three lines daily capacity of 27 pieces, adopts movable assembly table, track spacing of 6 meters, when the order batch is <50 pieces, one production line can be quickly removed and converted into a storage area, removal time <2 hours.

[0107] 3. Implementation of Physical Reconstruction Project

[0108] Dismantling and Sliding: The movable equipment of the original three mixed production lines will be slid to the new area along the pre-embedded steel plate track. The track has a load-bearing capacity of 5 tons, a sliding accuracy of ±5mm, and the track length covers the entire area of ​​production lines A1 to A3 (see...). Figure 2 (The track line is shown by the dashed line in the middle).

[0109] Zoning and isolation: The workshop is divided into dedicated work areas made of stainless steel using steel isolation barriers 211 (2.5m high) to avoid the mixing of welding shielding gases and the resulting carburization contamination of the welds;

[0110] Flexible configuration structure: The small-diameter area (A3) adopts a movable assembly table, and the track adopts a segmented pre-embedded design. Each 2-meter segment can be disassembled independently. When the order structure changes, the production line quantity B3 can be dynamically adjusted within 2 hours.

[0111] 4. Sensor 201 Deployment Labeling Instructions (Corresponding) Figure 2 (marked with a ☆)

[0112] ☆1 (Small Diameter Line A3): The laser displacement sensor 201 is installed on the cutting head of the blanking process M1, 1.2 meters away from the center line of the track, to avoid collisions with heavy pipe parts;

[0113] ☆4 (Small Diameter Line A3): The angle encoder and the pressure sensor 201 for the pipe bending process M2 are integrated into the base of the pipe bending mold. The RFID reader 202 is deployed at the entrance of the assembly process M3, 1.5 meters above the ground.

[0114] ☆7 (Small Diameter Line A3): The small diameter line uses a single sensor 201 to detect the flow rate of welding gas (☆7 is located at the branch point of the main gas pipeline), reducing equipment investment;

[0115] 5. Technical Effects and Solution Mapping

[0116] Stainless steel piping workshop 100: mold change time reduced from 35 minutes to 12 minutes / time, OEE increased from 58% to 81%;

[0117] Carbon steel pipeline workshop 100: The layout of this embodiment can be directly reused, only the protective gas is changed to a CO2+Ar mixture, and the tooling configuration parameters remain unchanged;

[0118] Copper Pipe Workshop 100: Add a preheating process before welding (about 5 minutes / pipe), and the production line quantity B1=B2=B3=1 line can meet the production capacity, and the total track length can be shortened by 30%.

[0119] Example 3: Edge Gateway Hard Real-Time Bottleneck Determination Logic

[0120] This embodiment details the implementation framework of the hard real-time control logic of claim 5, which conforms to the IEC 61131-3 standard and can be deployed without complete code.

[0121] 1. Determine the timing parameters of the process.

[0122] Sampling period: Δt = 30 seconds (the interval between Uᵢ and Qᵢ read from the PLC);

[0123] Single-time decision response: Total delay from reading to PLC triggering <1ms;

[0124] Cumulative verification period: 3×Δt=90 seconds (condition is met in 3 consecutive samplings);

[0125] 2. Core Judgment Steps

[0126] Data Acquisition: The edge gateway directly reads the real-time equipment utilization rate Uᵢ and the number of RFID202 work-in-process queues Qᵢ from the S7-1500 PLC via the Profinet protocol, with a reading latency of <0.5ms.

[0127] Conditional judgment: If Uᵢ>80% and Qᵢ(t)>Qᵢ(t-Δt) (the current queue is greater than the previous cycle queue), then M4_Counter=M4_Counter+1; otherwise, M4_Counter=0 and directly return to sampling wait.

[0128] Cumulative verification: When M4_Counter≥3, Mᵢ is determined to be the bottleneck process.

[0129] Hard real-time trigger: The edge gateway sets the memory register M4_Bottleneck_Flag=TRUE and immediately writes it to the PLC physical address DB100.DBX0.0 via the hard real-time industrial communication protocol to trigger the rate limit instruction. The write delay is <0.5ms.

[0130] Asynchronous Synchronization: In the next MQTT synchronization cycle (30 seconds), the M4_Bottleneck_Flag value is written to the Process_Node.Bottleneck_Flag field of the MES server for use in Kanban alarms and work order pushes.

[0131] 3. Hard Real-Time Communication Implementation Method

[0132] This embodiment uses the Profinet protocol (compliant with IEC 61158 standard) to achieve communication between the edge gateway and the PLC, with a total hard real-time response latency of <1ms. Those skilled in the art can use industrial Ethernet protocols of equivalent performance, such as EtherCAT or SercosIII, instead without affecting the core logic of this invention.

[0133] 4. Memory variable structure (edge ​​gateway VxWorks system)

[0134] M4_Utilization:REAL; / / Current device utilization rate (%);

[0135] M4_Queue:INT; / / Current number of RFID202 queues;

[0136] M4_Queue_Last:INT; / / Queue count in the previous cycle (historical value);

[0137] M4_Counter:INT; / / A counter for the number of consecutive increments;

[0138] M4_Bottleneck_Flag:BOOL; / / Process bottleneck status flag, 1=bottleneck, 0=normal.

[0139] 5. Timing Guarantee Description

[0140] The edge gateway has a built-in Cyclic_1ms real-time task, and the decision-making logic execution and PLC writing do not depend on the MES server or SQL database. Even if the MES fails, the edge layer can still independently complete bottleneck identification and equipment control, which conforms to the hierarchical real-time control architecture of claim 4.

[0141] Example 4: Value Stream Analysis and Waste Elimination

[0142] This example illustrates how to eliminate seven types of waste using value stream tools.

[0143] Current state value stream mapping (based on the VLGC project):

[0144] 7 processes: Material cutting process M1 → Pipe bending process M2 → Assembly process M3 → Welding process M4 → Visual inspection process M5 → X-ray inspection process M6 → Assembly and delivery process M7;

[0145] 7 Types of Waste Identification:

[0146] Waste due to waiting time: The average waiting time for materials before welding process M4 is 45 minutes (accounting for 11% of the cycle).

[0147] Wasteful handling: The cross-regional handling distance of work-in-process reaches 200 meters, taking 12 minutes per batch;

[0148] Wasted time: When changing molds, workers spend an average of 8 minutes each time searching for tools;

[0149] Waste of defective products: Mixed gases lead to a 18% weld rework rate;

[0150] Overproduction waste: Batch production results in an average of 18 units of work-in-process inventory.

[0151] Excessive processing waste: The M6 ​​inspection process in the radiography step is over-inspected with an excessive 30% sampling rate;

[0152] Inventory waste: Pipe inventory turnover days reach 45 days;

[0153] Optimization of future state value stream graph:

[0154] Establish a cyclical pull mechanism: The downstream process (assembly and delivery process M7) sends pull signals to the upstream process (material unloading process M1) through the MES Kanban board to achieve order-driven operation;

[0155] Implement a timed delivery system: AGVs (Automated Guided Vehicles) deliver every 30 minutes, with a delivery batch of ≤3 pieces, eliminating waiting and handling waste;

[0156] Elimination results: Total non-value-adding time decreased from 180 minutes to 45 minutes, with a waste elimination rate of 75%.

[0157] Example 5: Automatic Selection of Production Organization Method (including detailed calculation process)

[0158] This embodiment details the complete calculation process for determining the production organization method based on the processing batch n as described in claim 7.

[0159] Basic data preparation:

[0160] copy;

[0161] Total installation requirement R=309 units;

[0162] Small diameter R3 = 189 pieces;

[0163] Medium diameter R2 = 60 pieces;

[0164] Large diameter R1=60 pieces;

[0165] Production line configuration:

[0166] A1 (large diameter): B1 = 1 line, single line daily production capacity = 9 pieces;

[0167] A2 (medium diameter): B2 = 2 lines, daily production capacity of a single line = 9 pieces;

[0168] A3 (small diameter): B3 = 3 lines, daily production capacity of a single line = 9 pieces;

[0169] Basic production parameters:

[0170] Minimum batch single-stream job cycle = 90 minutes;

[0171] Small batch parallel sequential movement cycle = 160 minutes;

[0172] Maximum batch size limit: x=3;

[0173] Daily effective working time = 477 minutes;

[0174] S100 Steps: Theoretical Demand Batch Calculation and Production Line Bottleneck Identification

[0175] n3 = 189 / (9 × 3) = 7.0;

[0176] n² = 60 / (9 × 2) = 3.333;

[0177] n1 = 60 / (9 × 1) = 6.667;

[0178] The theoretical batch size of the system is n = max(7.0, 3.333, 6.667) = 7.0;

[0179] Bottleneck production line: A3 type small diameter production line (n3=7.0).

[0180] After completing the max(nᵢ) calculation, the MES server writes the bottleneck capacity production line identifier (A3 in this example) and the corresponding theoretical production batch nᵢ=7.0 into the Bottleneck_Line and Theoretical_Batch fields of the Production_Line table. This flag is only used for production planning decisions and delivery date baseline calculations, and does not trigger hard real-time control of the edge gateway. The edge gateway continues to independently execute equipment control based on process-level bottleneck determination.

[0181] Delivery period baseline = 7.0 days.

[0182] S200 Steps: Production Organization Decisions

[0183] Decision-making logic:

[0184] IF n=7.0≤1?→No;

[0185] x is the preset maximum batch size, x≥3, and the value is x=3.

[0186] IF n=7.0>x=3? → Yes → Use loop batch mode.

[0187] Loop batch parameters:

[0188] Number of cycles = 7 / 3 = 3 times;

[0189] The single cycle period Tx = 160 minutes;

[0190] Total production cycle Tz = 3 × 160 = 480 minutes;

[0191] The production cycle of A3 production line is 480 minutes, which is used as the delivery time benchmark for an order of R=309 pieces.

[0192] Configuration per batch:

[0193] Cycle 1: Produce 15 pieces (3 small + 1 medium + 1 large).

[0194] Cycle 2: Produce 15 pieces (3 small + 1 medium + 1 large).

[0195] Cycle 3: Produce 15 pieces (3 small + 1 medium + 1 large).

[0196] Production plan execution:

[0197] Daily cycle count = 477 / 160 ≈ 3 times (considering an 8-hour workday, approximately 477 minutes);

[0198] Daily output = 3 × 15 = 45 pieces;

[0199] Project duration = 309 / 45 ≈ 6.87 days ≈ 7 days.

[0200] Advantage verification:

[0201] Traditional sequential movement cycle = 90 × 7.0 = 630 minutes;

[0202] The optimization cycle for this system is 373.3 minutes.

[0203] Efficiency improvement = (630 - 373.3) / 630 = 40.7%;

[0204] Example 6: Application of Standard Formulas for Production Organization Methods

[0205] This embodiment illustrates the specific application of the sequential and parallel sequential movement formulas of claim 7.

[0206] Standard formula:

[0207] The formula for calculating the TOP production cycle of sequential movement is as follows:

[0208] ;

[0209] The formula for calculating the production cycle TOP for parallel sequential movement is:

[0210] ;

[0211] In the formula: x represents the batch; k represents the number of processes; ti represents the processing time of a single piece in the i-th process.

[0212] This system is used for:

[0213] One-time flow job (n≤1): using sequential movement, T=15+10+20+30+5+5+5=90 minutes;

[0214] Parallel sequential movement (n>1): Calculated using the formula, when x=3:

[0215] T=3×(15+10+20+30+5+5+5)-2×(min(15,10)+min(10,20)+...+min(5,5));

[0216] T = 3 × 90 - 2 × (10 + 10 + 20 + 5 + 5 + 5) = 270 - 2 × 55 = 160 minutes;

[0217] Formula meaning: Parallel sequential movement eliminates waiting time between processes through the min(tᵢ, ti+1) term, reducing the batch production cycle by 40.7%.

[0218] Example 7: Verification of the Overall Technical Effect of the System

[0219] This embodiment summarizes the synergistic technical effects of claims 1-8, and based on the VLGC project order for 309 stainless steel pipes, the improvement effects are as follows:

[0220] Indicator Item traditional methods This system Increase Technology source Production cycle 630 minutes 373.3 minutes 40.70% Claims 2+5+6 Bottleneck process identification accuracy 60% (labor) 100% 67% Claim 5 (Hard Real-Time) Identification response time 2 hours 90 seconds 98% Claim 5 (Edge Gateway) Equipment utilization rate 58% 79% 36% Claim 1 (S1 Reconstruction) Work-in-process inventory 18 items 4.5 pieces 73% Claims 5+6 mold change time 45 minutes 9 minutes 80% Claim 3 (Layered Acquisition) On-time delivery rate 72% 96.50% 34% Claim 2 (Cyclic Control)

[0221] Verification method: 15 orders were continuously run in the pipeline workshop of this plant (Q3 2024). The data was verified on-site by the classification society, and the MES system logs were exported from the Process_Node table.

[0222] Example 8: Implementation path for those skilled in the art

[0223] The hardware involved in this invention consists of industrial standard components, and the software logic conforms to the IEC 61131-3 standard. The implementation path is as follows:

[0224] Mechanical Engineer: Responsible for the design of sensor 201 mounting bracket, production line track laying, and layout of isolation barrier 211, based on the attached... Figure 2 Construction is underway, with a construction period of approximately 2 weeks.

[0225] Electrical Engineer: Configure the S7-1500 PLC using TIA Portal V17, import the function block (FB100_Bottleneck_Detection) provided by the edge gateway, and configure the Profinet network. The project will take approximately one week.

[0226] Automation Engineer: Deploy the Cyclic_1ms task in the Advantech WISE-Edge gateway, referring to the logical framework of Example 3, and bind it to the MES server via MQTT. The project time is about 3 days.

[0227] Total implementation time: approximately 3 weeks, no custom development required, all components are COTS (commercial off-the-shelf) products, within the scope of standard skills for mechanical / electrical engineers.

[0228] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A pipeline lean production management system and method, characterized in that, A hierarchical real-time control architecture is adopted, including: Edge computing gateways run hard real-time operating systems and execute millisecond-level control logic. The MES server is deployed at the management level, runs a general operating system, and performs second-level monitoring and historical analysis. The data acquisition module is deployed on the production equipment of the core processing process in the pipeline workshop (100) and connects the edge computing gateway and the MES server; The edge computing gateway communicates with the MES server and collaboratively executes the following steps: S1: The pipeline workshop (100) is physically reconstructed from a multi-material pipeline mixed production area into a single-material dedicated operation area, and divided into m types of flexible production lines {A1, A2, ..., A...} according to pipe diameter specifications. m The number of production lines of type i, Aᵢ, is Bᵢ. S2: Based on the real-time data from the data acquisition module, bottleneck processes are dynamically identified through a bottleneck determination algorithm. The algorithm is as follows: when the equipment utilization rate of a certain process is >80% and the number of work-in-process at its entry point continues to increase, the process is determined to be a bottleneck. S3: Based on the bottleneck identification results and the total installation requirement R, calculate the theoretical production batch nᵢ=Rᵢ / (single line capacity×Bᵢ), and automatically select the production organization method: if nᵢ≤1, adopt small batch one-pass flow operation; if nᵢ>1, adopt parallel sequential moving operation. S4: Generates control commands and sends them to the production line PLC via the edge computing gateway to execute production.

2. The pipeline lean production management system and method according to claim 1, characterized in that, The edge computing gateway also performs cyclic batch production control: When the theoretical production batch size nᵢ > x, a cyclic batch production mode is adopted, with each batch having a quantity of x and a production cycle of T. x ; Total production cycle Tz = (n / x)·T x The edge computing gateway automatically triggers the next batch production instruction after each batch is completed, without the need for manual intervention.

3. The pipeline lean production management system and method according to claim 1, characterized in that, The data acquisition module is configured hierarchically according to process type: Core process data acquisition unit: Sensors (201) and timers installed in the blanking process (M1) for detecting position, acquiring cutting length and processing time t1; sensors (201) installed in the pipe bending process (M2) for detecting angle and pressure, acquiring bending angle, forming pressure and processing time t2; sensors (201) installed in the welding process (M4) for detecting current and gas flow rate, acquiring welding parameters and processing time t4. Auxiliary process data acquisition unit: RFID (202) readers are deployed in the assembly process (M3), visual inspection process (M5), X-ray inspection process (M6), and assembly delivery process (M7) to record the start / end time t of the process by scanning the work-in-process pallet tags. 3、 t 5、 t 6、 t 7、 To monitor changes in the quantity of work-in-process; The workshop industrial Ethernet transmits the data from the aforementioned sensors (201) and RFID (202) to the edge computing gateway and MES server in real time.

4. The pipeline lean production management system and method according to claim 1, characterized in that, In the hierarchical real-time control architecture: The edge computing gateway maintains a lightweight state mapping table for millisecond-level bottleneck determination; The MES server stores a complete real-time state mapping model for second-level monitoring and historical backtracking. The real-time state mapping model includes: a real-time queue Qᵢ and a bottleneck marker register; The edge gateway and the MES server maintain model consistency through an asynchronous data synchronization protocol.

5. The system according to claim 1, characterized in that, The bottleneck determination algorithm executed by the edge computing gateway is as follows: The operating status of each process equipment is read from the data acquisition module according to the configurable sampling period Δt; The equipment utilization rate Uᵢ for process Mᵢ is calculated as: actual processing time / total observation time. When Uᵢ>80% and Qᵢ(t+Δt)>Qᵢ(t) for three consecutive sampling periods, Mᵢ is marked as a bottleneck process. Specifically, the edge computing gateway sets the bottleneck marker register of Mᵢ and directly triggers the PLC physical address through the hard real-time industrial communication protocol. After the MES server synchronizes the marker via asynchronous message, it triggers the Kanban audible and visual alarm and pushes the work order to the production supervisor's mobile terminal.

6. The pipeline lean production management system and method according to claim 1, characterized in that, In step S3 executed by the edge computing gateway, the bottleneck capacity identification and production batch determination method is as follows: The total installation requirements R given by the shipboard installation team are resolved into a set of corresponding requirements for various types of piping {R1, R2, ..., R...}. m }; Based on the number of production lines Aᵢ of type i, Bᵢ and their single-line capacity, calculate in real time the theoretical production batches nᵢ required to meet the various pipeline requirements Rᵢ, which is calculated as nᵢ=Rᵢ / (single-line capacity×Bᵢ). The theoretical production batch n of the system is determined to be the maximum value among all nᵢ, that is, n=max(nᵢ); The production line corresponding to the maximum value nᵢ is dynamically marked as the bottleneck capacity production line. Specifically, the bottleneck mark field of the production line is set in the SQL database of the MES server, and the production cycle of the bottleneck line is used as the production cycle benchmark for the pipeline workshop (100) to deliver the order requirement R.

7. The pipeline lean production management system and method according to claim 1, characterized in that, The logic for automatically selecting the production organization method in step S3 executed by the edge computing gateway is as follows: The theoretical production batch size of the calculation system is n = max(nᵢ). If n≤1, the edge computing gateway generates a single streaming job instruction, with a production cycle T. op This is the sum of the cycle times for each process. If n>1, the edge computing gateway generates parallel sequential movement instructions and dynamically calculates the optimal production cycle T based on the process cycle time. op .

8. The pipeline lean production management system and method according to claim 7, characterized in that, The parallel sequential movement instructions generated by the edge computing gateway include timing offsets for inter-process transfers: When the inter-process single-piece processing time tᵢ≥ti+1 is detected, the instruction embedding offset Δtᵢ=(x-1)·ti+1; The offset is used to precisely control the start time of the batch in the next process, avoiding idle waiting of the equipment.

9. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, performs the functions of the system as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program performs the functions of the system as described in any one of claims 1 to 8.

Images