© Strojni{ki vestnik 46(2000)1,5-13 © Journal of Mechanical Engineering 46(2000)1,5-13 ISSN 0039-2480 ISSN 0039-2480 UDK 658.563 UDC 658.563 Izvirni znanstveni ~lanek (1.01) Original scientific paper (1.01) Model strate{kega odlo~anja distribucije orodij z uporabo pravil mehke logike Model of Strategic Determination of Tool Distribution Based on the Fuzzy Rules Gorazd Vre~er - Franci ^u{ Cilj, ki ga želimo izpostaviti v prispevku, je izdelava modela mehke logike, s katerim simuliramo dinamični sistem distribucije orodij. Narejen je model gibanja orodij v proizvodnem procesu in skladišču srednje velikega podjetja. V podjetju je prilagodljivost proizvodnje v veliki meri odvisna od obsega zelo kakovostnih izdelkov, dobrega nadzora pretoka in strateške distribucije orodij, ki morajo biti na voljo vedno ob pravem času in v zadostnih količinah. Imamo dinamični pretok orodij, ki ga je mogoče nadzorovati z uporabo zakonov mehke logike in povratnih signalov. Z njimi primerjamo želene in dobljene vrednosti in krmilimo pretok. V prispevku je prikazan model, s katerim smo prikazali časovne odmike, ki nastanejo pri distribuciji orodij. Predstavljeni realni nadzorni sistem pretoka orodij v podjetu je sestavljen iz pretoka materiala, informacij in proizvodnih sredstev. V sistemu obstajajo zveze med posameznimi enotami podjetja, ki so povezana z informacijskimi tokovi, s katerimi smo povečali prilagodljivost sistema. © 2000 Strojniški vestnik. Vse pravice pridržane. (Ključne besede: pretok orodij, načrtovanje, modeli, logika mehka) This paper concerns the working out of a fuzzy algorithm model to simulate the dynamic system of tool distribution. A model of tool motion in the production process and in the store of a medium-size company is made. In a company, the production flexibility depends to a large extent on the high- quality manufacturing and the flow and strategic distribution of tools, which must always be available in time and in sufficient quantities. Tool flow is dynamical and can be monitored according to fuzzy algorithm principles by means of feedback signals with which the desired and obtained values are compared and the flow controlled. This paper shows the model with which we presented the time delays occurring during the tool distribution. The presented real tool distribution monitoring system of a company comprises the flow of materials, information and production means. The system features the interconnections between the individual units of the company, involved in the information flows with which the flexibility of the system was increased. © 2000 Journal of Mechanical Engineering. All rights reserved. (Keywords: tool distribution, planning, models, fuzzy algorithms) 0 UVOD Model prikazuje pretok orodja skozi skladišče v nekem proizvodnem podjetju. Večino orodja v proizvodnji porabimo, presežek orodja pa prodamo ali skladiščimo. Pretok orodja moramo optimizirati tako, da je količina orodja minimalna, zaradi česar zmanjšamo skladiščne stroške, obenem pa mora biti količina takšna, da ne pride do pomanjkanja, in bi nastali zastoji [4]. Model na sliki 1 je prikazan kot sistem s povratno zvezo, ki posreduje informacijski signal [6]. Ekonomske sisteme lahko opisujemo s pravili mehke logike, prav tako tudi tehnične sisteme ([7] in [10]). Naš primer je prikaz časovnega obnašanja zalog v skladišču, kjer nadzorujemo dotok orodja v skladišče 0 INTRODUCTION The model represents the flow of tools through the store of a production company. Most of the tools are consumed in production, the surplus tools are sold or stored. The tool flow must be optimized so that the quantity of tools is minimal, which re-duces the storage costs, but at the same time the quan-tity must be such that a tool shortage and resulting stoppages do not occur [4]. The model in Figure 1 shows a system with a feedback loop, which transfers an information signal [6]. The economic systems as well as the engineering systems can be described by fuzzy logic rules ([7] and [10]). Our example shows the time behavior of stocks in a store where the tool inflow into the store and isfFIsJBJbJJIMlSlCšD I stran 5 glTMDDC G. Vre~er - F. ^u{: Model strate{kega odlo~anja - Model of Strategic Determination in odtok iz njega, da bi zmanjšali stroške skladiščenja in proizvodnje. Z računalniškim programom smo izdelali sistem za simuliranje skladišča. Zahteve, ki jih moramo upoštevati pri nadzoru, so: 1. Pretok orodja v skladišču mora pokrivati lastno porabo, prodajo orodij in obvezno rezervo. 2. Uporabili bomo nadzorni sistem mehke logike, ki se uporablja pri odpravljanju motenj v sistemu. 3. V odvisnosti od pravil mehke logike, lahko izberemo obliko obnašanja sistema skladiščenja, lahko se pojavi recesija (pomanjkanje) ali konjunktura (presežek) orodij. 1 MODELIRANJE PROIZVODNIH SISTEMOV Z UPORABO ZAKONITOSTI MEHKE LOGIKE Vsak sistem, naj bo mehanski, elektronski ali organizacijski, je zapleten zaradi velikega števila odnosov med elementi, ki ga sestavljajo ([12] in [13]). Sistemski prijem je eden od načinov za iskanje najboljših ali vsaj dobrih rešitev problemov, ob upoštevanju celovitosti in omejitve sistema. Upravljanje sistemov je omogočeno z načelom povratne zveze [11], ki velja za katerikoli sistem, tudi biološki. Zaradi velike zapletenosti sistemov je najlaže dobiti celoten pregled s simuliranjem in modeliranjem. Lahko rečemo, da brez zmogljivih računalnikov upravljanje velikih zapletenih sistemov ne bi bilo mogoče [1]. Vsak sistem, ki ga obravnavamo, je omejen od okolja. V našem primeru, ko imamo skladišče orodja, predstavljajo okolje dobavitelji in porabniki. Elementi iz okolja so povezani prek vhodov v sistem in izhodov iz njega. Cilj obravnave je poiskati odnose v sistemu, kar je podlaga za simuliranje. Metode, ki jih bomo uporabljali, bodo temeljile na matematičnih zakonih in računalniških programih. Rezultati bodo grafično prikazani. Tehnični in tudi sistemi za nadzor pretoka orodij z uporabo pravil mehke logike temeljijo na načelu povratne zveze, pri katerem na sedanje stanje posameznega elementa vplivajo neposredno ali posredno prejšnja stanja. Sistem je sklenjen ali zančni sistem, pri katerem povratna zveza bistveno spreminja dinamiko sistema. Povratna zveza je lahko pozitivna ali negativna. Pozitivno zvezo poznamo v ekonomiji, biologiji, oziroma povsod, kjer se srečujemo z rastjo. Sistem brez povratne zveze ima takšno strukturo, da na delovanje ne vplivajo prejšnja stanja. Z modelom bomo analizirali gibanje zalog. V prispevku je predstavljeno nekaj zakonitosti mehke logike na primeru optimizacije zalog orodij v skladišču. Količina Q je tista kosovna količina orodij ki jo želimo imeti v skladišču. Da to dosežemo , moramo imeti pretok orodja q’, ki je odvod Q po času. Povečevanje količine orodja v skladišču, če nimamo odtoka, je prikazano na sliki 2 in enačbi (1), kjer je Vk varnostna količina. ^BSfiTTMlliC | stran 6 outflow from it are monitored in order to reduce the storage and production costs. By means of a computer program we made a system for the simulation of such a store. The following requirements must be taken into consideration for monitoring: 1. The tool flow in the store must cover the con-sumption, tool sales and an obligatory reserve. 2. A fuzzy algorithm monitoring system applicable for the removal of disturbances in the system will be used. 3. Depending on the parameters of the fuzzy algo-rithm it is possible to select the behavior of the storing system; recession (shortage) or upward trend (increase) of tools can occur. 1 MODELING OF PRODUCTION SYSTEMS BY MEANS OF FUZZY ALGORITHM RULES Any system, either mechanical, electronic or organizational is complex due to the large number of relationships between the elements of which it consists ([12] and [13]). A systematic approach is one of the ways of searching for the best, or at least satisfactory, solution to the problem by taking into account the completeness and limitations of the system. The management of systems is ensured by the feedback loop [11] applicable for any system, even biological. Because of the high complexity of the systems, an overall survey can be most easily obtained by means of simulation and modeling. It can be claimed that the management of large complex systems would not be possible without highly capable computers [1]. Any system to be dealt with is limited by the environment. In our case, when dealing with a tool store, the environment is represented by the suppliers and users. The elements from the environment are connected to the system through inputs and outputs. The aim of the treatment is to find the relationships in the system, which are the basis for the simulation. The methods, which will be used, are based on mathematical laws and software programs. The results will be graphically represented. The engineering and tool flow monitoring fuzzy algorithms for the management are based on the prin-ciple of a feedback loop, where previous states directly or indirectly influence the present state of the individual element. The system is a closed loop, where the feedback loop significantly changes the dynamics of the system. The feedback loop can be positive or negative. The positive feedback loop occurs in economics and biology, or anywhere where growth is encountered. The system without a feedback can have a structure where the previous states do not influence the functioning. By means of the model the stock movements will be analyzed. This paper presents some fuzzy algorithm rules in an example of optimization of tool stocks in a store. The quantity Q is the number of tools desired in the store. In order to achieve this, the tool flow must be q’, which is the derivative of Q by the time t. The increase of tools’ quantity in the store, if there is no outflow, is shown in Figure 2 and in equation (1) where Vk is the safety quantity. G. Vre~er - F. ^u{: Model strate{kega odlo~anja - Model of Strategic Determination • naročila / orders • lastna poraba / own consumption dobavitelji suppliers motnje disturbances Sl. 1. Model pretoka orodja skozi skladišče (prirejeno po [9]) Fig. 1. Model of tool flow through store(adapted from [9]) Q = q'-t +Vk(kos - piece) q'= 20 (kos/dan - piece/day) Vk = 40 (kos - piece) (1). PRETOK ORODJA TOOL FLOW varnostna količina safety tool quantity 10 15 čas time 25 dnevi days Sl. 2. Povečevanje količine orodij v skladišču pri konstantnem dotoku orodja, če ni porabe Fig. 2. Increase of tool quantity in store with constant tool inflow if there is no application Pri programiranju smo uporabili trapezno integracijo, ki je sicer manj natančna, vendar z njo dosežemo dobre rezultate, če zgostimo delitev. Prikaz integracije z integracijskim elementom je na sliki 3. For programming, the trapezoidal integration was used, which is less accurate, but by using it, good results are obtained if the division is high. The integration by means of an integration element is shown in Figure 3. G. Vre~er - F. ^u{: Model strate{kega odlo~anja - Model of Strategic Determination pretok orodij A (kosi/dan) tool flow (pieces/day) q’ X(tn+Dt) tn tn + Dt . čas (dnevi) t * time (days) Sl. 3. Integracija med dvema točkama, kjer površina pod krivuljo A pomeni količino orodja v skladišču Fig. 3. Integration between two points where the surface below curve A represents the tool quantity in the store Integracijo med dvema točkama poenostavimo s površino v obliki trapeza. Enačba površine A je: \dA= The integration between two points is simplified with the surface of a trapezoidal shape. Equation of surface A is: Xn+1)(tn+1 -tn) Na sliki 4 je prikazan simbol za integracijski element z več vhodi in enim izhodom. V sistem vstopajo orodja, pri čemer se njihove vrednosti seštevajo. V odvisnosti od tehnološkega procesa lahko z njim ponazorimo časovno odvisno gibanje orodij v različnih orodnih sistemih. Pri modeliranju uporabimo program EXCEL, ki ima vgrajene matematične module. Nestabilnost popravimo s spremembo parametrov mehke logike. Načelo je uporabljeno tudi v skladiščih, kjer nadzorujemo pretok orodja. (2). 2 Figure 4 shows the symbol for the integration element with several inputs and one output. Depending on the technological process, it is possible to represent the tool motion dependent on time in different tool systems. For modeling, the EXCEL program, incorporating mathematical modules, is used. The instability is corrected by using the fuzzy algorithm. Such a principle is used also in a store where the tool flow is supervised. STRUŽENJE TURNING VRTANJE DRILLING FREZANJE MILLING . rezalna ploščica /cutting insert • držalo /holder svedri / drills vpenjalni del / clamping part steblo / stem glava / head valjasto frezalo / cylindrical milling cutter steblasto frezalo / stem milling cutter kotno frezalo / angular milling cutter vpenjalo / clamping device frezalne glave / milling heads rezalne ploščice / cutting insert MODULNA orodja MODULAR tools (VARIOLOCK) sistemsko orodje s stožcem ISO system tool with ISO cone . osnovnovno držalo / basic holder . adapterji / adapters . rezalni modul / cutting module rezalne ploščice / cutting inserts ORODJE TOOLq trni / mandrels • puše s stožcem morse / bushes with morse orodni sistem tool system cone frezalne glave / milling heads Sl. 4. Orodni sistemi za različne tehnologije, ki jih ponazorimo z integracijskimi elementi Fig. 4. Tool system for different technologies, represented with integration elements grirMsfcflMISDSD VH^tTPsDDIK stran 8 G. Vre~er - F. ^u{: Model strate{kega odlo~anja - Model of Strategic Determination Osnovna naloga vodenja in nadzora je hitra stabilizacija sistema in doseg želene vrednosti. V primeru, da se pojavijo v sistemu motnje (sprememba dotoka ali odtoka orodij), mora nadzor omogočiti takojšen popravek in ponovno stabilizacijo sistema. Z računalniškim nadzorom laže načrtujemo pretok orodij za oskrbovanje večjega števila strojev [1]. V prispevku je prikazan model oskrbe strojev z orodij s pomočjo dobaviteljev in skladišča orodij. Zahteve po orodjih so odločilni parametri za načrtovanje porabe. Model je preverjen na praktičnem primeru [8], izdelan je v programskem jeziku EXCEL, s čimer smo nazorno dokazali pravilnost simuliranja. Dana je še varnostna količina orodij. Transportne in skladiščne stroške želimo zmanjšati, pri čemer nam pomaga program, ki izračuna stroške ob spremenjenih parametrih. Namen modeliranja je poiskati povezave med posameznimi enotami podjetja, med poslovnimi enotami, med proizvodnjo, nabavo in prodajo. Vsak model mora omogočiti uporabniku, da izvaja odločitve in možnost preskusa celotnega poteka. V vseh dinamičnih sistemih, kjer nadziramo pretok z algoritmi mehke logike, je pomemben povratni tok informacij [2]. Referenčni model pomaga pri zgradbi organizacije sistema. S simuliranjem se korakoma približujemo realnemu stanju. Pri tem uporabljamo bazo podatkov, ki jo moramo nenehno dopolnjevati, obenem pa so osnovni podatki za naš organizacijski sistem. Na sliki 5 je prikazan zahtevan pretok Yr orodja skozi skladišče v nekem časovnem roku. V skladišču imamo orodja preveč ali premalo (recesija, konjunktura).V skladišču je zaposleno več ljudi, ki skrbijo za nemoten pretok in za nastavljanje orodja. Prepustnost skladišča je odvisna od frekvence dotoka orodij in hitrosti nastavljanja, skladiščenja in transporta orodij. Izhod orodja iz skladišča Y je amplitudno enak vhodu v skladišče X, le da je med njima časovni razmik. Nastavitev, skladiščenje in transport povzročajo časovne premike med količino orodij, ki vstopa v skladišče in izstopa iz njega. The basic task of the management and supervision is fast stabilization of the system and achieving the desired value. In the case of distur-bances occuring in the system (change of tool in-flow and outflow) the supervision must ensure im-mediate correction and re-stabilization of the system. By means of computer monitoring the tool flow to a great number of machines is more easily planned [1]. This paper describes the model of tool supply to machines by tool suppliers. The requirements for tools are decisive parameters for the planning of consumption. The model is verified on a practical example [8]; it is made in the EXCEL program language so that the correctness of the simulation can be clearly demonstrated. A safety quantity of tool is also anticipated. Reduction of transport and storing costs is wished for; to this end a program is used that calculates the costs with changed parameters. The purpose of modeling is to find interconnections between the individual units of a company, between the business units, production, purchasing and sales. Each model must enable the user to implement the decisions and the possibility of testing the entire process. In all the dynamic systems where the tool flow is controled by fuzzy algorithms, the return flow of information is important [2]. The reference model helps us to build the organization of the system. By means of simulation approach the real state we step by step. To this end we use a base of data, which must be constantly up-dated and at same time are the basic data for our organizational system. Figure 5 shows the required tool flow, Yr, through the store within a certain time. Increasing and decreasing the number of tools in the store (recession, upward trend) are shown. Several employees, taking care of trouble-free flow and adjustment of tools, are employed in the store. The freedom of movement through the store depends on the frequency of the tool inflow and the speed of adjusting, storing and transporting the tools. The tool outflow from the store, Y, with respect to the amplitude, is equal to the inflow into the store, X, except that there is a time delay between them. Adjustments, storing and transport cause the time delays between the tool quantities that enter and leave the store. Yr Ym recesija recession Yr trdi prehod hard transition Ym konjunktura a Yr upward trend mehki prehod "soft transition Sl. 5. Pretok orodij v skladišču v odvisnosti od parametrov logično mehkega krmilnika Fig. 5. The trends of tools in store are obtained with dependence on the fuzzy controller ^vmskmsmm stran 9 G. Vre~er - F. ^u{: Model strate{kega odlo~anja - Model of Strategic Determination 2 LOGIČNO MEHKI KRMILNIK ZA KRMILJENJE PRETOKA ORODIJ Na splošno je logično mehki krmilnik zelo primeren za sistem krmiljenja z več spremenljivkami. Znano je, da ima krmilnik z dvema vhodoma in enim izhodom dovolj dobre karakteristike za krmiljenje pretoka orodij ([3] in [5]). Dva vstopna signala v logično mehki krmilnik se imenujeta: napaka pretoka E in sprememba pretoka CF. Napaka pretoka E je razlika med referenčnim (priporočenim) pretokom Yr in dobljenim pretokom Ym. Sprememba pretoka CF pa je razlika med dobljenim pretokom i-tega koraka in (i-1)-tega koraka. Obe enačbi zapišemo: 2 FUZZY CONTROLLER FOR TOOL DISTRIBUTION CONTROL Generally, the fuzzy controller is very suitable for the multi-variable control system. In experiments, it was found that a two inputs and one output fuzzy controller is suitable to obtain good performance in tool quantity control ([3] and [5]). The two inputs of the fuzzy controller are the output error, E, and the output change of tools, CF. The flow error, E, is the difference between the reference quantity of tools, Yr, and the real quantity, Ym . The tool quantity change, CF, is the difference between the real quantity of the ith period and the quantity of tools of the (i – 1)th period. That is: E =Yr– Ym (i) CF = Ym (i) – Ym (i – 1) (3), (4). Izhod nadzora mehke logike X je definiran kot sprememba ukaza o pretoku orodij. Zaradi tega je ukaz o spremembi pretoka i-te periode X(i) enak seštevku spremembe ukaza pretoka i-te periode dX(i) in ukaza o spremembi pretoka (i-1)-tega periode X(i-1). Enačba se glasi: The output of the fuzzy controller, X, is de-fined as the change of the tool quantity command. Therefore, the tool quantity command of the ith output period, X(i), is equal to the summation of the change of the tool quantity command of the ith period, dX(i), and the tools quantity command of the (i – 1) ith period, X(i– 1). That is: X(i) = X(i – 1) + dX(i) (5). 3 KRMILNI ALGORITEM MEHKE LOGIKE IN PRAVILA Realno številčno stanje je treba realno stanje prikazati s pojmi mehke logike. Kot rezultat lahko navedemo naslednje jezikovne nize: NV večja negativna vrednost NS srednja negativna vrednost NM manjša negativna vrednost NČnič PV večja pozitivna vrednost PS srednja pozitivna vrednost PM manjša pozitivna vrednost Niza za vhod in izhod logično mehkega krmilnika sta bila načrtovana v območju [-1,1]. Potrebno je določiti stopnje vstopnih in izstopnih vrednosti logično mehkega krmilnika. Slika 6 prikazuje pripadnostne funkcije na izstopu iz logično mehkega krmilnika, kar je ukaz za spremembo količine orodij dX, prikazane v enačbi (5). Rezultat je: ukaz za določanje količine orodij se ne spremeni, če je majhna razlika med izstopno količino orodij in referenčno količino, ki se ustvari v odvisnosti od motnje. 3 FUZZY CONTROL ALGORITHMS AND RULES In order to let the quantity of the real world be fuzzified to the quality of the fuzzy world, it is necessary to qualify the quantitative statement. As a result, the following linguistic sets are assigned: NB negative big, NM negative medium, NS negative small, ZE zero, PB positive big, PM positive medium, PS positive small. The fuzzy sets for the input and output of the fuzzy controller have both been designed in the range of [-1, 1]. In other words, the scaling of the input and output values of the fuzzy controller is nec-essary. Figure 6 shows the shape of membership func-tions for the output of the fuzzy controller, i.e. the change of the tools quantity command, dX, see equa-tion (5). As a result, the tools quantity command is unchanged when a small variation between the output tool quantity and the reference quantity is generated owing to a disturbance. grin^SfcflMISDSD VBgfFMK stran 10 G. Vre~er - F. ^u{: Model strate{kega odlo~anja - Model of Strategic Determination NV-NB NS-NM NM-NS NC-ZE PV-PB PM-PS -1 -2/3 -1/3 0 1/3 2/3 1 Sl. 6. Pripadnostne funkcije Fig. 6. Shape of membership functions . Ym čas / time a1 b1 c1 d1 Sl. 7. Oasovni odziv količine orodij Fig. 7. Tool quantity response in time interval Slika 7 prikazuje primer količine orodij na izstopu iz modela. Za točko a1 je napaka količine E enaka PV in sprememba količine CF je N. Zaradi tega je v točki a1 izhod krmilnega signala mehke logike dX enak PB. Glede na prejšnje opise, lahko zapišemo stanja v naslednjih oblikah: - če je E enak PB in CF enak ZE sledi, da je dX enak PB. Podobno zapišemo v točkah b1, c1, d1,..., naslednja stanja: - če je E enak N? in CF enak PV sledi da je dX enak NS - če je E enak NB in CF enak NČ sledi da je dX enak NV - če je E enak NČ in CF enak NV sledi da je dX enak PS Končno lahko zapišemo določila mehke logike v odločitveni preglednici, kakor je prikazano v preglednici 1. Figure 7 shows an example of the tool quantity response. For the point a1 the quantity error, E, is PB and the quantity change, CF, is ZE. Therefore, at the point a1, the output of the fuzzy controller, dX, should be PB. Therefore, based on the above discussion, a fuzzy conditional statement can be described in the following form: - if E is PB and CF is ZE then dX is PB. Similarly, at the points b1, cl, d1,..., a number of fuzzy conditional statements can also be described as: - if E is ZE and CF is PB then dX is NM - if E is NB and CF is ZE then dX is NB - if E is ZE and CF is NB then dX is PM Finally, for convenience, these fuzzy con-ditional statements have been formulated and ex-pressed as a decision table, see Table 1. Preglednica1. Logično mehki krmilni ukazi za spremembo količino orodij Table 1. Fuzzy control rules for the tool quantity change CF E NV/NB NS/NM NM/NS NČ/ZE PM/PS PS/PM PV/PB NV/NB NV/NB NV/NB NV/NB NV/NB NV/NB NV/NB NV/NB NS/NM NS/NM NS/NM NV/NB NV/NB NV/NB NV/NB NV/NB NM/NS NM/NS NM/NS NM/NS NS/NM NS/NM NV/NB NV/NB NČ/ZE PS/PM PM/PS PM/PS NČ/ZE NM/NS NM/NS NS/NM PM/PS PV/PB PS/PM PM/PS PM/PS NČ/ZE NM/NS NM/NS PS/PM PV/PB PV/PB PS/PM PS/PM PM/PS NČ/ZE NM/NS PV/PB PV/PB PV/PB PV/PB PV/PB PS/PM PM/PS NČ/ZE ^vmskmsmm stran 11 G. Vre~er - F. ^u{: Model strate{kega odlo~anja - Model of Strategic Determination 4 SIMULIRANJE PRETOKA ORODIJ MED PROIZVODNIM PROCESOM 4 SIMULATION OF TOOLS DISTRIBUTION DURING THE MANUFACTURING Poraba orodij brez logično mehkega krmiljenja Usage of tools without fuzzy logic control 1200 1000 800 600 400 200 0 časovne enote (dnevi) working time period (days) Sl. 8. Količina orodij v skladišču med proizvodnim procesom Fig. 8. The tool quantity in the storage during manufacturing Opis modela: V začetku proizvodnega procesa imamo 1000 kosov orodij. Količina mora zadostovati za oskrbo proizvodnega procesa. Slika 8 prikazuje, kako se količina orodij zmanjšuje. Določili smo varnostno količino, ki prepreči pomanjkanje orodij. Varnostna količina je 300 kosov. Ko upade količina pod to vrednost, je treba naročiti od dobaviteljev novo orodje. Uporabo orodij določimo statistično in je predstavljena v enačbi kot naključno število RND. Za preprečitev pomanjkanja moramo pripraviti krmiljenje, temelječe na pravilih mehke logike. Enačbi (6) in (5) prikazujejo model porabe orodij, kjer smo z i označili časovne periode. Description of model: In the beginning there are 1000 tool pieces. This quantity must cover all necessary consumption during the production process. In Figure 8 we find that the quantity of tools during the production process has been reduced. We have to determine the safety quantity, which prevents a tool shortage. The safety quantity is 300 tool pieces. If the tool quantity is under this level then we must supply supplementary tools from suppliers to the magazine. The usage of tools is statistically determined and is represented as random function RND. For preventing a tool shortage we must prepare control based on fuzzy rules. The equations (6) and (5) represent the model of tool consumption where i is the time period. Ym (i) = Ym (i -1)-Ym (i -1). RND • 0,1 + X(i) (6). Na sliki 9 vidimo odvisnost števila kosov orodij pri krmiljenju, temelječem na pravilih mehke logike. Varnostna količina 300 kosov je tista optimalna količina, ki zmanjšuje stroške. In Figure 9 we can see the tool quantity controled using fuzzy rules. The safety quantity of 300 pieces is the optimum quantity also reducing the tool costs. Poraba orodij z logično mehkim krmiljenjem Usage of tools usage with fuzzy logic control 1000 800 600 400 200 0 časovne enote (dnevi) working time period (days) Sl. 9. Krmiljenje količine orodij, temelječe na pravilih mehke logike Fig.9. Control of tools‘ quantity based on fuzzy rules grin^sfcflMISDSD VH^tTPsDDIK stran 12 G. Vre~er - F. ^u{: Model strate{kega odlo~anja - Model of Strategic Determination 5 SKLEP 5 CONCLUSION Pretok orodij je dinamičen, saj je odvisen od časa in se prilagaja zakonitostim mehke logike. Pojavijo se amplitudni in fazni premiki, ki jih ponazorimo z zakonitostmi mehke logike. Minimalna finančna sredstva za material so eden od bistvenih dejavnikov prilagodljive proizvodnje. Za dosego prilagodljive proizvodnje je treba analizirati vse komponente, ki vplivajo na proizvodni proces. V skladu s strategijo podjetja morajo biti dobavitelji. Ti omogočajo hitro, kakovostno in poceni dobavo orodij in drugih materialnih sredstev, ki sodelujejo v procesu. Optimalno količino smo določili s statističnimi metodami, kjer moramo upoštevati obvezno rezervo. Koristna je praktična uporaba zakonitosti mehke logike pri modeliranju. The tool flow is dynamical since it depends on time and adapts itself to fuzzy algorithm rules. Amplitude and phase delays occur that are represented by the use of fuzzy algorithm rules. Minimum funds for tools are one of key factors in flexible manufacturing. To achieve flexible manufacturing it is necessary to analyze all the compo-nents participating in the production process. Also, the suppliers must conform to the company’s strategy. They assure a fast, cheap and high-quality supply of tools and other material means participating in the process. The optimum quantity was determined by means of statistical methods where it is necessary to take into account the obligatory reserve. The use of fuzzy algorithm rules in modeling of the tool flow proves that they are a useful aid also for monitoring the tool distribution. 6 LITERATURA 6 REFERENCES [I] Katalinič, B., S. Celar (1996) Einsatz der Computersimulation für die Auslegung und Steuerung von FFS. Proceedings of DAAM Symposium, Editor: Katalinic B., ISBN 3-901509-02-X, Vienna, 203-204. [2] Warnecke, G., Augustin, H., Förster, H., Rauch, C, N. Sepet (1996) Aufbau und Anwendungen eines integrierten Prozeßmodells fur die Produktion. Industrie Management 12 5, 21-25. [3] Tarng, Y.S. and Y.S. Wang (1994) A new adaptive controler for constant turning force. Int J Adv Manuf Technol 9:211-216, 1994 Springer-Verlag London Limited [4] Čuš, F., J. Balič (1996) Rationale Gestaltung der organisatorischen Abläufe im Werkzeugwesen. International Conference on Computer Integration Manufacturing Zakopane, May 14-17. [5] Nestler, A., D. Fichtner (1998) An approach to technological database-possibilities for the determination of cutting values with neural networks. Proceedings, Flexible Automaton and Intelligent Manufacturing, 577-586. [6] Brown, M. (1986) Differential equations and their applications. 3rd edition, Springer Verlag, New York, Heidelberg Berlin. [7] Fikri Dweiri (1999) Fuzzy development of crisp activity relationship charts for facilities layout. Computers & Industrial Engineering 36, 1-16. [8] Vrečer, G, F. Čuš (1998) Planning of tool flow in fexible production. V. Dobrzanski, Leszek A. (ur.). Proceedings of the 7th International Scientific Conference Achievements in Mechanical & Materials Engineering AMME’98, Gliwice-Zakopane, Poland. [9] Oertli, PC. (1977) Praktische Wirtschaftskybernetik. Carl Hansen Verlag, ISBN 3-446-12234-6, München Wien, 1-434. [10] Tzung-Pei Hong & Tzung-Nan Chuang (1999) A new trianglar fuzzy Johnson algorithm, Computers & Industrial Engineering 36, 179-200. [II] Zakovorotny, V. L., Bordatchev, E. V., TS. Sankar (1997) Variational formulation for optimal multi-cycle deep drilling of small holes. Journal of Dynamic System, Measurement, and Control, Vol.119, 553-560. [12] Rober, S.J., Shin, Y., D.D.I. Nwokah (1997) A digital robust controller for cutting force control in the end milling process. Journal of Dynamic System, Measurement, and Control, Vol.119,146-152. [13] Katalinic B., E. Fabian (1996) Automatische Identifikation als Schnittstelle für die Synchronisierung von Material- und Informationsfluss. Computer Integration Manufacturing, Zakopane, May 14-17, 185-192. Naslov avtorjev: dr. Gorazd Vrečer prof.dr. Franci Čuš Fakulteta za strojništvo Univerze v Mariboru Smetanova 17 2000 Maribor Authors’ Address:Dr. Gorazd Vrečer Prof.Dr. Franci Čuš Faculty of Mechanical Engineering University of Maribor Smetanova 17 2000 Maribor, Slovenia Prejeto: Received: 16.8.1999 Sprejeto: Accepted: 29.2.2000