Production optimization method and system based on workshop key in-process monitoring deduction after order production scheduling

Abstract

The invention belongs to the technical field of automobile manufacturing, and particularly discloses a production optimization method and system based on workshop key in-process monitoring deduction after order production scheduling, and the method comprises the following steps: taking an actual time point as a collection instantaneous time point, and automatically collecting in-process data of each order of a workshop; setting production rhythm data of each line body in a future production day to obtain a pre-production plan; presetting the maximum value and the minimum value of a working point based on the working data of each sequence and the pre-production plan, comparing with each working point of the pre-production plan, judging whether the placement of each working point of the pre-production plan is reasonable or not, and if the working point is not reasonable, feeding back the result to the production scheduling system to formulate a reasonable production plan; data in the production process are analyzed in real time, abnormal conditions and potential problems are detected, and if the abnormal conditions are detected, an early warning mechanism is triggered. According to the technical scheme, transparent in-process deduction is provided, in-process changes of joints are obtained, and a reasonable production plan is made.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of automobile manufacturing, and relates to a production optimization method and system based on order production scheduling and post-production workshop key in-process monitoring deduction. BACKGROUND

[0002] With the rapid development of the automobile industry, consumers' demand for automobiles is becoming stronger and stronger. The manufacturing process of automobiles mainly includes stamping, welding and coating, assembly and logistics, and these links are closely linked, so various problems are inevitable in large-scale production activities.

[0003] With the intensification of competition in the automobile market, the market response speed and informatization level of automobile factories are increasingly required, and enterprises must quickly respond to market changes and order demand in the production process, and make real-time adjustments to the production plan.

[0004] Currently, most of the automobile industry does not deduce the results after production scheduling, and the in-process deduction is limited to the execution state of the workshop, and there is a lack of monitoring of the future execution plan after the in-process change of the workshop is not refined.

[0005] During the post-production scheduling and workshop execution process, the storage workshop with key in-process storage has the problems of insufficient in-process stop line and excessive in-process, which affects production efficiency and increases costs. SUMMARY

[0006] The purpose of the present application is to provide a production optimization method and system based on order production scheduling and post-production workshop key in-process monitoring deduction, to solve the problem of insufficient in-process stop line and excessive in-process of the storage workshop with key in-process storage during the post-production scheduling and workshop execution process.

[0007] In order to achieve the above purpose, the basic scheme of the present application is: a production optimization method based on order production scheduling and post-production workshop key in-process monitoring deduction, comprising the following steps:

[0008] The actual time point is used as the collection instantaneous time point, and the in-process data of each sequence of the workshop is automatically collected;

[0009] According to the order delivery period and the production efficiency of the comprehensive material host factory, the production rhythm data of each line body of the future production day is set, and the pre-production plan is obtained;

[0010] Based on the in-process data of each sequence and the pre-production plan, the maximum and minimum values of the in-process point are preset, and compared with each in-process point of the pre-production plan, to determine whether the placement of each in-process point of the pre-production plan is reasonable, and if the in-process is not reasonable, the result is fed back to the production scheduling system to develop a reasonable production plan;

[0011] Real-time analysis of data during the production process is performed to detect anomalies and potential problems. If an anomaly is detected, an early warning mechanism is triggered.

[0012] The working principle and beneficial effects of this basic solution are as follows: This technical solution provides transparent work-in-process simulation, which can clearly obtain the changes in work-in-process at each stage every hour after the production plan is issued and executed, avoiding waste due to excessive work-in-process and production line stoppage due to insufficient work-in-process. It formulates reasonable production plans and monitors them in real time, triggering an early warning mechanism when anomalies occur, rearranging the production plan and outputting an early warning signal.

[0013] Furthermore, the work-in-process data for each sequence in the workshop includes:

[0014] Workshops: Welding workshop, WT (Welding), Painting workshop, TA (Painting), Final assembly workshop, workshop i ∈ {stamping, welding, painting, final assembly}, workstation j ∈ {1..M} i};

[0015] The data includes regional data for the work-in-process (WBS) buffer area at the connection points between the welding workshop and the painting workshop, and the work-in-process (PBS) buffer area at the connection points between the painting workshop and the final assembly workshop, as well as data for brand, vehicle model, and configuration dimensions: CT(i,j,t): Cycle Time of workstation (i,j) at time t, CAP(k): Maximum capacity of buffer k, α(i,j,t): Equipment availability, β(i,j,t): Quality pass rate, Q(o,t): Release quantity of order o at time t, and WIP(k,t): Quantity of work-in-process in buffer k at time t.

[0016] State equation: WIP(k,t+Δt)=WIP(k,t)+∑inflow-∑outflow

[0017] The inflow and outflow are determined by the rhythm and path.

[0018] Collect the necessary data for easy use.

[0019] Furthermore, based on order delivery dates and the overall production efficiency of the main equipment manufacturer, the production cycle time data for each production line on future production days is set, and the pre-production plan is obtained as follows:

[0020] Convert the APS plan and real-time device data into List E:

[0021] Inputs: APS scheduling results, real-time orders, equipment status:

[0022] Generation: Event queue E = {e1, e2, ...}, event types include: order arrival, workstation changeover, fault, quality inspection rework, specifically:

[0023] ② Order arrival: Time = APS release time,

[0024] ② Changeover: Time = Plan Start - Changeover Duration,

[0025] ③ Failure: Time = Last Repair + Random MTBF,

[0026] ④ Rework: Time = Finish + Random Rework Duration;

[0027] Event-driven simulation

[0028] Using DES, clock advance function:

[0029] T_next = min{finish_time(e) | e E}

[0030] Where, T_next: the global earliest event trigger time that the simulation clock will advance to next; min{}: the mathematical minimum function; finish_time(e): the event trigger time field carried in event e; e: a single event instance in the event queue E; E: the pending event priority queue maintained by the current simulation engine;

[0031] Calculation logic: take the minimum value of the trigger time of all events in the event queue E, which is the time T_next that the global clock will jump to next: first traverse E, take out the time field of each event, take the minimum value of these times, and let T_next = the minimum value; the simulation clock directly jumps to T_next and executes the corresponding event; Derivation index calculation:

[0032] Workshop i average tact Takt i :

[0033]

[0034] Where, ∑ t∈H : sum over all discrete sampling points t in the statistical period H, H: statistical interval; j: workstation number in workshop i; M i : total workstation number or parallel device number in workshop i; CT(i,j,t): nominal process cycle time of workstation (i,j) at time t; α(i,j,t): instantaneous device availability of workstation (i,j) at time t, ∑ t∈H ∑ j CT(i,j,t)·α(i,j,t): sum of nominal process cycle availability time of all workstations (i,j) in workshop i at time t;

[0035] Numerator: accumulate the "effective tact" CT·α of each sampling point t and each workstation j to get the total effective processing time of workshop i in H;

[0036] Denominator: |H|·M _i This represents "total number of sampling points × total number of workstations", which is the total number of possible processing cycles.

[0037] Ratio: Gives the weighted cycle time required to produce one product on average throughout the entire workshop and the entire statistical interval;

[0038] Workshop i Output i :

[0039]

[0040] Among them, Output i : The total number of orders that are finally completed and rolled off the production line in workshop i within the statistical period H; finish(o): The final completion timestamp of order o in workshop i, o: a single production order;

[0041] During the production simulation, all orders o are iterated and summed to obtain ∑_{o}. The final completion time finish(o) of workshop i is checked. If finish(o) falls within the interval H, the count is incremented by 1; otherwise, it is ignored. The accumulated result is the actual output of the workshop within H. i ;

[0042] Maximum WIP in buffer k:

[0043]

[0044] Among them, MaxWIP k The maximum number of work-in-progress items in buffer k within the statistical period H; Take the maximum value for all discrete sampling times t within time period H; WIP(k,t): the instantaneous work-in-process quantity of buffer k at time t;

[0045] During the simulation process, whenever the inventory in buffer zone k changes (vehicle entry / exit), WIP(k,t) is updated immediately. The WIP(k,t) values ​​of all sampling points are recorded, and the maximum value is taken as MaxWIP. k ;

[0046] Line stop risk index:

[0047]

[0048] Among them, Risk i : The downtime risk index of workshop i within the statistical period H, ∑ k∈i : Sum all buffers k belonging to workshop i; MaxWIP k: Peak value of buffer k in H, in units of product, CAP(k): Maximum physical capacity of buffer k, 0.9: Risk alert line coefficient; MaxWIP k / CAP(k): Peak utilization rate of buffer k;

[0049] When calculating in-process, for each buffer k, first calculate the peak utilization rate MaxWIP k / CAP(k), subtract 0.9: if utilization ≤ 90%, ≤ 0 in parentheses, ReLU output 0, no risk contribution; if utilization > 90%, output positive difference, add up the positive differences of all buffers in workshop i to get Risk i Example: A factory workshop i has 3 buffers:

[0050] ① Buffer A: MaxWIP a = 190, CAP a = 200 → 190 / 200 - 0.9 = 0.05

[0051] ② Buffer B: MaxWIP β = 180, CAP β = 200 → 180 / 200 - 0.9 = 0.00

[0052] ③ Buffer C: MaxWIPc= 210, CAPc= 200 → 210 / 200 - 0.9 = 0.15

[0053] Risk i = ReLU(0.05) + ReLU(0.00) + ReLU(0.15) = 0.05 + 0 + 0.15 = 0.20, if the system is set τ = 0.1, Risk i > τ, trigger red alert.

[0054] According to the order delivery date and the production efficiency of the main material host factory, set the production pace data of each line body in the future production day, get the pre-production plan.

[0055] Further, make a reasonable production plan:

[0056] Optimal in-process = initial in-process + transferred-in vehicles - transferred-out vehicles = in-process value setting interval.

[0057] Simple operation, easy to use.

[0058] Further, real-time analysis of data in the production process is carried out to detect abnormal situations and potential problems. If abnormal situations are detected, the early warning mechanism is triggered, specifically:

[0059] Input inference indicators: {Takt i , Outputi MaxWIP k Risk i};

[0060] Objective function is:

[0061]

[0062] Where MaxWIP k : the peak number of in-process products in buffer k within H,

[0063] Takt i : the time for workshop i to produce one in-process product;

[0064] Risk i : the risk index for workshop i to produce in-process products;

[0065] CAP(k): the maximum physical capacity of buffer k;

[0066] Output i : the actual output of workshop i within statistical period H;

[0067] Demand i : the planned demand of workshop i within the same statistical period H;

[0068] γ: the upper limit coefficient of buffer safety;

[0069] W i is the weight of the i-th workshop;

[0070] λ is the over-limit punishment weight coefficient of in-process products;

[0071] Decision variables:

[0072] ΔTakt i : the takt adjustment amount of workshop i,

[0073] ΔCAP k : the virtual upper limit adjustment amount of buffer k,

[0074] Analytical formula:

[0075]

[0076] Where η, ξ are learning rates, γ is a safety threshold, 0 < γ < 1, and ε is a small amount of zero prevention;

[0077] Update rule:

[0078] Takt i ← Takt i + ΔTakt i

[0079] CAP(k)←CAP(k)+ΔCAP k

[0080] Convergence judgment:

[0081] If |Output i –Demand i | / Demand i <σ and MaxWIP k <γ·CAP(k),

[0082] Where, Output i : Actual output of workshop i in the statistical period H, Demand i : The planned demand (scheduling target) of workshop i in the same statistical period H; For all buffer zones k; MaxWIP k : Peak WIP of buffer zone k; γ: Buffer safety upper limit coefficient; CAP(k): Maximum physical capacity of buffer zone k; σ: Yield deviation tolerance threshold;

[0083] Judgment logic:

[0084] Output side: |Output i –Demand i | / Demand i <σ, that is, the actual output and the planned demand deviation is within ±σ is considered to meet the standard;

[0085] Buffer side: MaxWIP k <γ·CAP(k), that is, all buffer zones peak inventory do not reach the dangerous water level (γ times capacity) is considered safe;

[0086] Only when the two conditions are met at the same time, the "derivation convergence" is judged, and the outer ring APS rearrangement is not triggered; If any condition is not met, the rearrangement or adjustment is started, and a warning signal is output.

[0087] Real-time analysis of data in the production process, detection of abnormal conditions and potential problems, if abnormal conditions are detected, trigger the warning mechanism, and guarantee production.

[0088] The application also provides a production optimization system based on order scheduling workshop key in-process monitoring deduction, comprising a data acquisition unit and a processing unit, the data acquisition unit is used for collecting each sequence in-process data of the workshop, and transmitting to the processing unit;

[0089] The processing unit executes the method of the application, carries out the monitoring and deduction of the workshop key in-process, and completes production optimization.

[0090] The system is based on a data acquisition unit and a processing unit, carries out the monitoring and deduction of the workshop key in-process, and completes production optimization. BRIEF DESCRIPTION OF DRAWINGS

[0091] Figure 1 is a flowchart of the production optimization method based on the monitoring and deduction of the workshop key in-process after order production scheduling of the application;

[0092] Figure 2 is a chart of the data of each sequence in-process of the workshop of the production optimization method based on the monitoring and deduction of the workshop key in-process after order production scheduling of the application;

[0093] Figure 3 is a chart of the data of each line body production rhythm of the future production day of the production optimization method based on the monitoring and deduction of the workshop key in-process after order production scheduling of the application. DETAILED DESCRIPTION

[0094] The embodiments of the application are described in detail below, examples of which are shown in the drawings, wherein the same or similar reference signs represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are only used to explain the application, and cannot be understood as a limitation of the application.

[0095] In the description of the application, it should be understood that the terms "longitudinal", "transverse", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the application and simplifying the description, and therefore cannot be understood as indicating or implying that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the application.

[0096] In the description of the application, unless otherwise specified and limited, it should be noted that the terms "mounting", "connection" and "connection" should be understood broadly, for example, it can be mechanical connection or electrical connection, or the communication between two elements, or direct connection, or indirect connection through an intermediate medium, and those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0097] The application discloses a production optimization method based on the monitoring and deduction of the workshop key in-process after order production scheduling, which provides transparent in-process deduction, can clearly let the planner know the change of each joint in-process every hour after the execution of the production plan, avoid waste caused by too much in-process, and avoid line stop caused by too little in-process.Figure 1 As shown, the production optimization method based on order production scheduling and workshop key in-process monitoring deduction includes the following steps:

[0098] The actual time point is taken as the collection instantaneous time point, and the workshop in-process data of each sequence is automatically collected; the in-process data of each sequence in the workshop is the data of each line, WBS, PBS and other areas, and the brand, vehicle type, configuration and other dimensions can be counted according to the requirements, such as Figure 2 As shown.

[0099] According to the order delivery period and the production efficiency of the comprehensive material host factory, the production rhythm data of each line on the future production day is set, and the pre-production plan is obtained; according to the order delivery period and the production efficiency of the comprehensive material host factory, the production rhythm data of each line on the future production day is set, such as Figure 3 As shown.

[0100] Based on the in-process data of each sequence and the pre-production plan, the maximum and minimum values of the in-process points are preset, and compared with each in-process point of the pre-production plan, whether the placement of each in-process point of the pre-production plan is reasonable is judged, if the in-process is not reasonable, the result is fed back to the production scheduling system, and a reasonable production plan is made; after the in-process data of each sequence is collected, the production data of the key points is maintained, the in-process deduction and monitoring can be realized through the operation and statistical method of the pre-production plan combined with the basic line production data, and the key production data can be adjusted to achieve the optimal production efficiency, and assist the artificial decision.

[0101] The data in the production process is analyzed in real time, and abnormal conditions and potential problems are detected, if the abnormal condition is detected, the early warning mechanism is triggered.

[0102] Based on the actual in-process data extraction after the production shift, the plan queue is calculated after the production plan is made, and the calculation result can be displayed in the form of a chart. The user can maintain different basic data according to the deduction result to meet the optimal production in-process and the highest efficiency.

[0103] In a preferred scheme of the application, the in-process data of each sequence in the workshop includes:

[0104] Workshop: welding workshop, W-T, painting workshop, T-A, assembly workshop, workshop i∈{stamping, welding, painting, assembly}, station j∈{1..M i};

[0105] The data includes regional data for the work-in-process (WBS) buffer area at the connection points between the welding workshop and the painting workshop, and the work-in-process (PBS) buffer area at the connection points between the painting workshop and the final assembly workshop, as well as data for brand, vehicle model, and configuration dimensions: CT(i,j,t): Cycle Time of workstation (i,j) at time t, CAP(k): Maximum capacity of buffer k, α(i,j,t): Equipment availability, β(i,j,t): Quality pass rate, Q(o,t): Release quantity of order o at time t, and WIP(k,t): Quantity of work-in-process in buffer k at time t.

[0106] State equation: WIP(k,t+Δt)=WIP(k,t)+∑inflow-∑outflow

[0107] The inflow and outflow are determined by the rhythm and path.

[0108] In a preferred embodiment of the present invention, the method for obtaining the pre-production plan by setting the production cycle data of each production line on future production days based on the order delivery date and the overall production efficiency of the main equipment manufacturer is as follows:

[0109] Convert the APS plan and real-time device data into List E:

[0110] Inputs: APS scheduling results, real-time orders, equipment status:

[0111] Generation: Event queue E = {e1, e2, ...}, event types include: order arrival, workstation changeover, fault, quality inspection rework, specifically:

[0112] ⑤ Order arrival: Time = APS release time

[0113] ② Changeover: Time = Planned Start - Changeover Duration

[0114] ⑥ Fault: Time = Last Repair + Random MTBF

[0115] ⑦ Rework: Time = Completion time + Random rework duration;

[0116] Event-driven simulation

[0117] DES is simulated using discrete events, with the clock advance function as follows:

[0118] T_next=min{finish_time(e)|e∈E}

[0119] T_next = min{finish_time(e) | e E} where T_next: the next global earliest event trigger time that the simulation clock will advance to; min{}: the mathematical minimum function; finish_time(e): the event trigger time field carried in event e; e: a single event instance in the event queue E; E: the pending event priority queue maintained by the current simulation engine;

[0120] Calculation logic: take the minimum value of the trigger time of all events in the event queue E, that is, the time T_next that the next global clock will jump to: first traverse E, take out the time field of each event, take the minimum value of these times, let T_next = the minimum value; the simulation clock directly jumps to T_next and executes the corresponding event; derivation index calculation:

[0121] Average Takt of workshop i i :

[0122]

[0123] where ∑ t∈H : sum all discrete sampling points t in the statistical period H, H: statistical interval; j: workstation number in workshop i; M i : total workstation number or parallel device number in workshop i; CT(i,j,t): nominal process cycle of workstation (i,j) at time t; α(i,j,t) instantaneous device availability of workstation (i,j) at time t, ∑ t∈H ∑ j CT(i,j,t)·α(i,j,t): sum of nominal process cycle availability time of all workstations (i,j) in workshop i at time t;

[0124] Numerator: add up the "effective Takt" CT·α of each sampling point t and each workstation j to get the total effective processing time of workshop i in H;

[0125] Denominator: |H|·M i , which represents "total number of sampling points × total number of workstations", that is, the total number of possible processing Takt;

[0126] Ratio: gives the weighted Takt time required to produce an average product in the entire workshop and the entire statistical interval;

[0127] Output of workshop i i :

[0128]

[0129] where Output iTotal number of orders (or vehicles) finished at workshop i in period H, finish(o): final completion timestamp of order o at workshop i, o: individual production order (contains attributes such as vehicle model, quantity, process path, etc.); H: statistical interval, such as a shift, a day;

[0130] In the production simulation process, all orders o are traversed and summed to obtain ∑ i ;

[0131] Maximum WIP of buffer area k:

[0132]

[0133] Where, MaxWIP k : Maximum number of in-process products in buffer area k in period H; Take the maximum value for all discrete sampling time points t in period H; t: sampling time (clock advancement point); H: statistical interval, such as a shift, a day; WIP(k, t): instantaneous number of in-process products in buffer area k at time t;

[0134] In the simulation process, whenever the inventory of buffer area k changes (vehicle enters / leaves the warehouse), update WIP(k, t) immediately, record all the values of WIP(k, t) at the sampling points, and take the maximum value, i.e. MaxWIP k ;

[0135] Risk of line stop index:

[0136]

[0137] Where, Risk i : Risk of line stop index of workshop i in period H, ∑ k∈i : Sum all buffer areas k belonging to workshop i; ReLU(x) linear rectifier function, MaxWIP k : Peak number of in-process products in buffer area k in H, CAP(k): Maximum physical capacity of buffer area k, 0.9: risk alert line coefficient (90% capacity threshold); MaxWIP k / CAP(k): Peak utilization rate of buffer area k;

[0138] When calculating in-process, for each buffer area k, first calculate the peak utilization rate MaxWIP k / CAP(k), minus 0.9: if utilization ≤ 90%, the value in the bracket is ≤ 0, ReLU outputs 0, no risk contribution; if utilization > 90%, output positive difference value, add the positive difference value of all buffer zones in workshop i to obtain Risk i

[0139] Example: a factory workshop i has 3 buffer zones:

[0140] ① Buffer A: MaxWIP a = 190, CAP a = 200 → 190 / 200 - 0.9 = 0.05

[0141] ② Buffer B: MaxWIP β = 180, CAP β = 200 → 180 / 200 - 0.9 = 0.00

[0142] ③ Buffer C: MaxWIPc= 210, CAPc= 200 → 210 / 200 - 0.9 = 0.15

[0143] Risk i = ReLU(0.05) + ReLU(0.00) + ReLU(0.15) = 0.05 + 0 + 0.15 = 0.20, if the system is set τ = 0.1, Risk i > τ, trigger red early warning.

[0144] Preferably, a reasonable production plan is made:

[0145] Work-in-Process in production = initial Work-in-Process + vehicles transferred in - vehicles transferred out = Work-in-Process value set interval.

[0146] In a preferred scheme of the present application, real-time analysis is performed on data in the production process to detect abnormal conditions and potential problems, and if an abnormal condition is detected, an early warning mechanism is triggered, specifically:

[0147] Input inference indicators: {Takt i , Output i , MaxWIP k , Risk i};

[0148] The objective function is:

[0149]

[0150] Wherein, MaxWIP k : the peak number of Work-in-Process in buffer zone k in H,

[0151] Takt iTime of workshop i to produce one work-in-process;

[0152] Risk i Risk index of workshop i to produce work-in-process;

[0153] CAP(k): Maximum physical capacity of buffer k;

[0154] Output i Actual output of workshop i in statistical period H;

[0155] Demand i Planned demand of workshop i in the same statistical period H;

[0156] γ: Upper limit coefficient of buffer safety;

[0157] W i Weight of the i-th workshop;

[0158] λ is the weight coefficient of work-in-process over-limit penalty, and the specific weight, upper limit and risk index can be set in advance; Decision variables:

[0159] ΔTakt i Takt adjustment amount of workshop i,

[0160] ΔCAP k Virtual upper limit adjustment amount of buffer k,

[0161] Analytical formula:

[0162]

[0163] where η, ξ are learning rates, γ is a safety threshold, 0 < γ < 1, and ε is a small amount to prevent zero;

[0164] Update rule:

[0165] Takt i ←Takt i +ΔTakt i

[0166] CAP(k)←CAP(k)+ΔCAP k

[0167] Convergence criterion:

[0168] If |Output i –Demand i | / Demand i <σ and MaxWIP k <γ·CAP(k),

[0169] wherein Output i : actual output of workshop i in the statistical period H, Demand i : planned demand (scheduling target) of workshop i in the same statistical period H; holds for all buffers k; MaxWIP k : peak work-in-process quantity of buffer k; γ: buffer safety upper limit coefficient; CAP(k): maximum physical capacity of buffer k; σ: yield deviation tolerance threshold;

[0170] Judgment logic:

[0171] Yield side: |Output i –Demand i | / Demand i <σ, that is, the actual output and the planned demand deviate within ±σ and are considered to meet the standard;

[0172] Buffer side: MaxWIP k <γ·CAP(k), that is, the peak inventory of all buffers does not reach the dangerous water level (γ times the capacity) and is considered to be safe;

[0173] Only when both conditions are met, it is determined that "derivation converges", and the outer ring APS rearrangement is not triggered; if either condition is not met, rearrangement or adjustment is started, and a warning signal is output.

[0174] The application also provides a production optimization system based on order scheduling and workshop key in-process monitoring derivation, comprising a data acquisition unit and a processing unit, the data acquisition unit is used to collect various in-process data of the workshop and transmit to the processing unit.

[0175] The processing unit executes the method described in the application to monitor and derive the key in-process of the workshop, and complete the production optimization.

[0176] The application monitors various key in-processes, realizes production progress control, can infer subsequent pre-execution plans according to the existing line in-process situation, can directly observe whether the placement of each in-process point is reasonable, formulates a reasonable production plan, avoids excess and insufficient in-process in the production process, discovers and solves problems in production in time, reduces downtime, and thus improves production efficiency.

[0177] The monitoring of various key in-processes analyzes and mines production data, finds out bottlenecks and problems in the production process, provides a basis for improving the production process, provides data analysis and decision support, and thus reduces costs.

[0178] Each key in-process monitoring realizes comprehensive monitoring and management of the production process through real-time collection and processing of data from the production site. The production process is transparent and controllable.

[0179] Each key in-process monitoring reduces unnecessary in-process laying, shortens the product manufacturing cycle, and improves the delivery efficiency of orders through reasonable in-process laying.

[0180] The key in-process monitoring system detects abnormal conditions and potential problems through real-time analysis of data in the production process based on pre-set rules and algorithms. Once an abnormal condition is detected, the early warning mechanism is triggered immediately to notify relevant personnel to handle the abnormality.

[0181] Each key in-process monitoring can timely discover problems in the production process through real-time collection and monitoring of data in the production process, thereby improving production quality.

[0182] Automobile order production line body in-process product deduction and monitoring of each production process helps to improve the overall production management level and enhance the market competitiveness of enterprises.

[0183] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the described specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0184] Although the embodiments of the present application have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and purposes of the present application, and the scope of the present application is defined by the claims and their equivalents.

Claims

1. A production optimization method based on order scheduling and subsequent workshop critical work-in-process monitoring and simulation, characterized in that, Includes the following steps: The actual time point is used as the instantaneous time point for collection, and the data of each process in the workshop is automatically collected; Based on order delivery dates and the overall material OEM production efficiency, we set the production cycle data for each production line on future production days to obtain a pre-production plan. Based on the work-in-process data and pre-production plan, the maximum and minimum values ​​of the work-in-process points are preset and compared with each work-in-process point in the pre-production plan to determine whether the placement of each work-in-process point in the pre-production plan is reasonable. If the work-in-process is unreasonable, the result is fed back to the production scheduling system to formulate a reasonable production plan. Real-time analysis of data during the production process is performed to detect anomalies. If an anomaly is detected, an early warning mechanism is triggered.

2. The production optimization method based on order scheduling and post-scheduling critical work-in-process monitoring and simulation as described in claim 1, characterized in that, The work-in-process data for each sequence in the workshop includes: Workshops: Welding workshop, WT (Welding), Painting workshop, TA (Painting), Final assembly workshop, workshop i ∈ {stamping, welding, painting, final assembly}, workstation j ∈ {1..M} i }; The data includes regional data for the work-in-process (WBS) buffer area at the connection points between the welding workshop and the painting workshop, and the work-in-process (PBS) buffer area at the connection points between the painting workshop and the final assembly workshop, as well as data for brand, vehicle model, and configuration dimensions: CT(i,j,t): Cycle Time of workstation (i,j) at time t, CAP(k): Maximum capacity of buffer k, α(i,j,t): Equipment availability, β(i,j,t): Quality pass rate, Q(o,t): Release quantity of order o at time t, and WIP(k,t): Quantity of work-in-process in buffer k at time t. State equation: WIP(k,t+Δt)=WIP(k,t)+∑inflow-∑outflow The inflow and outflow are determined by the rhythm and path.

3. The production optimization method based on order scheduling and post-scheduling workshop key work-in-process monitoring and simulation as described in claim 2, characterized in that, Based on order delivery dates and the overall production efficiency of the main equipment manufacturer, the following method is used to determine the production cycle time data for each production line on future production days and obtain the pre-production plan: Convert the APS plan and real-time device data into List E: Inputs: APS scheduling results, real-time orders, equipment status: Generation: Event queue E = {e1, e2, ...}, event types include: order arrival, workstation changeover, fault, quality inspection rework, specifically: Order arrival time: = APS release time Changeover: Time = Planned Start - Changeover Duration Fault: Time = Last Repair + Random MTBF Rework: Time = Completion + Random rework duration; Event-driven simulation, using Discrete Event Simulation (DES), clock advance function: T_next=min{finish_time(e)|e∈E} Where, T_next: the earliest global event trigger time that the simulation clock will advance to next; min{}: the mathematical minimum function; finish_time(e): the event trigger time field carried in event e; e: a single event instance in event queue E; E: the priority queue of pending events maintained by the current simulation engine; Calculation logic: The minimum trigger time of all events in event queue E is taken as the next global clock jump time T_nex. First, E is traversed, the time field of each event is extracted, and the minimum value among these times is taken, setting T_next = this minimum value. The simulation clock directly jumps to T_next and executes the corresponding event. Derivation index calculation: Workshop i Average Cycle Time Takt i : Where, ∑ t∈H : Sum of all discrete sampling points t within the statistical time interval H, where H: statistical interval; j: workstation number within workshop i; M i : Total number of workstations or parallel equipment in workshop i; CT(i,j,t): Nominal process cycle of workstation (i,j) at time t; α(i,j,t): Instantaneous equipment availability rate of workstation (i,j) at time t, ∑ t∈H ∑ j CT(i,j,t)·α(i,j,t): The sum of the nominal process cycle movable time of all workstations (i,j) in workshop i at time t; Molecular: During the deduction, the "effective cycle time" CT·α of each sampling point t and each workstation j is accumulated to obtain the total effective processing time of workshop i within H; Denominator: |H|·M _i This represents "total number of sampling points × total number of workstations", which is the total number of possible processing cycles. Ratio: Gives the weighted cycle time required to produce one product on average throughout the entire workshop and the entire statistical interval; Workshop i Output i : Among them, Output i : The total number of orders that are finally completed and rolled off the production line in workshop i within the statistical period H; finish(o): The final completion timestamp of order o in workshop i, o: a single production order; During the production simulation, all orders o are iterated and summed to obtain ∑_{o}. The final completion time finish(o) of workshop i is checked. If finish(o) falls within the interval H, the count is incremented by 1; otherwise, it is ignored. The accumulated result is the actual output of the workshop within H. i ; Maximum WIP in buffer k: Among them, MaxWIP k The maximum number of work-in-progress items in buffer k within the statistical period H; Take the maximum value for all discrete sampling times t within time period H; WIP(k,t): the instantaneous work-in-process quantity of buffer k at time t; During the extrapolation process, whenever the inventory of buffer k changes within time period H, WIP(k,t) is updated immediately. The WIP(k,t) values ​​of all sampling points are recorded, and the maximum value among them is taken as MaxWIP. k ; Line stop risk index: Among them, Risk i : The downtime risk index of workshop i within the statistical period H, ∑ k∈i : Sum all buffers k belonging to workshop i; MaxWIP k : Peak work-in-process quantity in buffer k within H; CAP(k): Maximum physical capacity of buffer k; 0.9: Risk warning threshold coefficient; MaxWIP k / CAP(k): Peak utilization of cache k; During the computation, for each cache k, the peak utilization MaxWIP is first calculated. k / CAP(k), minus 0.9: If utilization ≤ 90%, the value within parentheses ≤ 0, ReLU output 0, no risk contribution; if utilization > 90%, output a positive difference value, summing the positive differences of all buffers in workshop i to obtain Risk. i .

4. The production optimization method based on order scheduling and post-scheduling critical work-in-process monitoring and simulation as described in claim 1, characterized in that, To formulate a reasonable production plan: Optimal inventory = Initial inventory + Transfer-in vehicles - Transfer-out vehicles = Inventory value setting range.

5. The production optimization method based on order scheduling and post-scheduling critical work-in-process monitoring and simulation as described in claim 2, characterized in that, Real-time analysis of data during the production process is performed to detect anomalies and potential problems. If an anomaly is detected, an early warning mechanism is triggered, specifically: Input projection metrics: {Takt i Output i MaxWIP k Risk i }; The objective function is: Among them, MaxWIP k : Peak number of work-in-progress items in buffer k within H, Takt i The time it takes for workshop i to produce one work-in-process item; Risk i Risk index of work-in-process production in workshop i; CAP(k): The maximum physical capacity of buffer k; Output i : The actual output of workshop i during the statistical period H; Demand i : The planned demand of workshop i within the same statistical period H; γ: Cache safety limit coefficient; W i Let be the weight of the i-th workshop; λ is the weighting coefficient for penalties for exceeding work-in-process limits; Decision variables: ΔTakt i Workshop cycle time adjustment amount ΔCAP k : Adjustment amount of the virtual upper limit of the cache k Analysis formula: Where η and ξ are the learning rates, γ is the safety threshold, 0 < γ < 1, and ε prevents the elimination of zero small quantities; Update rules: Tact i ←Tact i +ΔTact i CAP(k)←CAP(k)+ΔCAP k Convergence criterion: like and Among them, Output i : Actual output of workshop i during the statistical period H, Demand i : The planned demand of workshop i within the same statistical period H; This holds true for all buffers k; MaxWIP k : Peak work-in-process quantity in buffer k; γ: Buffer safety limit coefficient; CAP(k): Maximum physical capacity of buffer k; σ: Production deviation tolerance threshold; Judgment logic: Production side: | Output i –Demand i | / Demand i <σ, meaning that the deviation between actual output and planned demand is within ±σ, is considered to meet the standard; Cache side: In other words, it is considered safe if the peak inventory of all buffer zones has not reached the danger level; If both conditions are met, the deduction is considered to have converged, and there is no need to trigger the outer loop APS rearrangement; if either condition is not met, the rearrangement or adjustment will be initiated and an early warning signal will be output.

6. A production optimization system based on order scheduling and post-scheduling critical work-in-process monitoring and simulation, characterized in that, It includes a data acquisition unit and a processing unit. The data acquisition unit is used to collect in-process data from each stage of the workshop and transmit it to the processing unit. The processing unit executes the method described in any one of claims 1-5 to perform monitoring and simulation of key work-in-process in the workshop and complete production optimization.

Images