Gj=lF(j)+LTij For online control, Gj for each tool is calculated for each round and the tool with the smallest Gj is selected. For all types of sister knives, the above methods can be used to match and complete the control of the entire tool flow. Mean control of tool flow optimization examples and results Hydraulic Flange Caps,Hydraulic Pump Adapter,Hydraulic Hose Protector,Hydraulic Tank Accessories Dongguan Jinzhuang Hydraulic Technology Co., Ltd. , https://www.jinzhuanghydraulic.com
1 Mean control Mean control is the overall average selection method for the elements to be optimized. The basic idea is that there is an ordered element A={a1, a2, ..., an}, and the sequence of elements arranged is
Face=symbol>a=a(1) a(2)...a(k),ai∈A,i∈[1,k]; sorting element a △∈A, unsorted element sequence b=b(1) b(2)...b(p), b(j)∈A,j∈[1,p], p+k=n-1.
There is an evaluation function F(a, aΔ, b, C), where C is (
Face=symbol>a,aΔ,b) and other objects' functions. due to
The total number of face=symbol>a and b arrangements is the factorial multiple of the total number of elements they contain, and the objective function F is a function of the arrangement of a and b.
The number of elements of face=symbol>a and b is slightly larger. To find the optimal F and its a, aΔ, and b in an exhaustive manner, there is a problem that the amount of calculation is too large. If you put
The face=symbol> elements of a and b and their effect on F are treated like
Face=symbol>a is only composed of k same "average" elements, b is composed of p same "average" elements, then F is only considered by what elements a, a?, b are composed of, and it is not necessary to consider Its arrangement. Select all the elements in b as a Δ, find one of the best F, and the corresponding a Δ is fixed. And so on, you can arrange the elements of b in turn.
It can be shown that this algorithm is a polynomial algorithm.
2 Control mechanism of tool flow 2.1 Evaluation function In FMS, parts are formed according to the processing sequence of multiple parts and the process requirements of each part, thereby selecting different types of tools. There can be more than one tool of the same type. The choice of sister tools with exactly the same processing functions is the subject of this article.
Although the sister tools have the same processing functions, their service life is not necessarily the same. It is always hoped that each tool will do its best to reduce the life of the tool. According to this goal, an evaluation function F is established. In the part-to-tool-requirement sequence, a is taken as the set of already ranked elements, a Δ is the element that is being tested, and b is the set of elements that are not arranged. 
In the formula
Gj - the average remaining life of the jth sister tool, which is divisible by ti 
After the remainder
Ti - the remaining service life of the j-th sister tool - The average use time of each demand in the sequence of the unsorted parts for the tool
Τi - Demand time of the ith sequence in the sequence of unordered parts to the tool 
p - the total number of unsorted parts to the tool, ie the length of b, i∈b
2.2 Recursive Calculation Process
Face=symbol>b-aΔ1, find F(1) using equations (1) and (2) and write down, then replace aΔ1
In face=symbol>b, choose another element in b as aΔ2.
Face=symbol>b′=b-aΔ2, find F(2) and write down. According to this method, p elements in b are tested once as aΔ, until F(p) is obtained, and a minimum is selected from F(*), fix the order of aΔ*, and use aΔ* as the last element of a, then make a new round of aΔ selection from the remaining b until all the elements in b are finished.
When comparing and selecting F(*), there are sometimes multiple equal minimum F(i), you can choose one of them, or you can take the opportunity to fine-tune: If you want to use the remaining life is focused on a small number of tools, select F If the residual life span of the gj is expected to be spread over a large number of tools, the F of gj with a small mean square error is selected.
3 Global decision 3.1 Tool lag time Tool lag time is the time for the machine tool to stop waiting for the tool, which directly affects the production efficiency of the FMS. Tool control should try to minimize tool lag time. Qualitatively speaking, the position of the tool at the time of the machine's need determines the time of the tool's hysteresis. The time's size and position relation is: location in other machine tools> local tool magazine in other machine tools> in the central tool magazine> in the local Partial magazine.
Suppose LTij is the delay time for the jth tool waiting for the i-th machine tool to stop, Pk is the remaining processing time for the part when the tool is machining the kth machine tool part, and tpkj is the use position of the tool j from the other kth machine tool. The transfer time to the local local magazine, takj is the transfer time of the tool j from the other kth machine's local magazine to the local local magazine, and tcj is the transfer time of the tool j from the central magazine to the local local magazine. Then there is LTij=Pk+tpkj (when tool j is at the position of the k-th machine)
LTij=takj (when tool j is in the k-th machine's partial magazine) LTij=tcj (when tool j is in the center magazine)
The above equation ignores the tool up/down time between the upcoming position of the tool and the local local magazine.
3.2 The total evaluation function of the total evaluation function The total evaluation function of the tool flow should consider not only the utilization of the tool but also the high production efficiency of the FMS. therefore
Discover through these examples:
(1) In the calculation of a selected tool, if the evaluation functions F of a plurality of tools are all equal, and they are the minimum value, the useless life remaining is used.
The tool with the largest mean square deviation of gj, the total tool life utilization rate obtained is obviously better than the service life of one tool with the smallest F, and the unused remaining life of the former is concentrated to a few tools, which is beneficial for later use. . The table shows the results of the first method. However, in addition to using the service life, the lag time of the tool should also be considered. This goal sometimes contradicts the first method. What method is adopted depends on the specific circumstances.
(2) The calculation time of the algorithm increases with the increase in the number of parts requiring tools and is slower than the increase in the number of tools. The main factor affecting the calculation time is the number of tools.
(3) Both Example 4 and Example 5 used 30 sister knives, the number of times the tool was required was 76 and 200, and the calculation time was 36 seconds. It is relatively rare to have as many as 30 sister tools installed in a system in FMS. Therefore, the calculation time of the algorithm can fully meet the requirements.
(4) The total tool life utilization ratio is the ratio of the total tool use time to the total tool life. The examples all reach over 90%.
FMS's tool management control employs an averaging control algorithm to have a satisfactory effect on optimization aimed at reducing the tool's useless remaining life.
This method chooses a Δ from a global perspective and has a large possibility to find the global optimal route. The calculation method is simple and suitable for on-line or off-line control or optimization.
FMS tool flow online optimization control strategy
In the tool management of FMS, the tool flow path decision is the most critical and the most complex control problem. In addition to ensuring the correct selection of tools, it also relates to whether it can reduce the time for loading and unloading tools and maximize the use of tool life. Reduce production costs and increase production efficiency. However, the complexity of the problem is high. The current optimization control mainly adopts a heuristic algorithm, the calculation is cumbersome, and it is difficult to grasp the optimization effect. In addition, there are still many areas to be improved in the multi-objective and on-line control methods. This article proposes a mean optimization control strategy to solve the above problems.