200424052 玖、發明說明 【發明所屬之技術領域】 方、本發明係一種以模穴壓力為基礎之射出成型品質監控 :法尤==穴壓力_:進f射出成型… 有放辨識其成σπ缺陷特徵之監控方法。 【先前技術】 按’、塑膠材料具有輕巧、财腐姓、成本低廉、絕緣性 土及易成型加工等優點，而可作為目前各種產業大量製造 …勺原料’其中’在眾多塑膠加工方法中，係以射出成 1 1為、9遍適用，且由於塑膠材料在加熱塑化後，具非 t 體之特性，故在射出成型機將溶融之塑料射入成型 、，中’此期間為一段無法透視之流動成型過程，因 引述充填模穴的過程中，其操作者若無妥善之品質管 复▲ 私払，而於射出成型條件中盲目以試誤法來調整 /、蒼數’則對於整，射 射出成i作業而S，無疑會造成時間 土膝原料成本的浪費。 而*見針對射出成型品質所提出之監控方法，主要係200424052 发明 Description of the invention [Technical field to which the invention belongs] The present invention is a kind of injection molding quality monitoring based on cavity pressure: Fayou == cavity pressure_: injection molding ... It can be identified that it is a σπ defect Method of monitoring characteristics. [Previous technology] According to ', plastic materials have the advantages of lightness, wealth, low cost, insulating soil, and easy forming and processing, etc., and can be used in a large number of various industries at present ... spoon raw materials' among them' among many plastic processing methods, It is suitable for injection 9 times, and the plastic material has the characteristics of non-t after heating and plasticizing. Therefore, the injection molding machine will melt the plastic into the injection molding. The perspective of the flow molding process, because the process of filling the cavity is quoted, if its operator does not have proper quality control, ▲ private, and in the injection molding conditions, the trial and error method is used to adjust blindly. , Injection into i operation and S, will undoubtedly cause a waste of time and raw material costs. * See the monitoring method proposed for injection molding quality, which is mainly
( Statistic Process control,SPO 3于位置、射嘴壓力、或模穴壓力實施監控，其中，由於熔 :塑料人模穴㈣，其直接接觸之物體即為模壁本身， *右方、杈壁表面埋設訊號感測器，則為最直接且可最 ::‘革k烙膠在杈穴内之流動狀態與壓力變化，進而可馮 精較精確之數值作為調整參數或監控品f之依據，故在^ 把成效上仍以模穴壓力進行品質監控為最適當之方法 4S4 5 200424052 穴壓力之品質監控方法’雖可針對射出成型品質之 員：進：監控’惟該方法在監控的過程中，仍須藉由操 、 断監控成品之品質，因而、止出人六次、、/5 、 乍 無法對其成品之缺陷特性： “、的浪費’且亦 特 毛邊或短射等做更進一步的 " …、法有效確保射出成型後之成品品質。 、 【發明内容】 =此’為改善丽述傳統模穴壓力之品質監控方法之 』’本發明者乃研發設計出一種 、 出成型品質監控方法」，好為基礎之射 U取即渴人力賁源的方式，針對 ^成型品質之良劣進行監控，並可有效辨識其成品 ^特徵，是為其主要之發明目的。 、 =了可達到前述的發明目的’本發明所運用的技術手 ;在於提供—種以模穴屋力為基礎之射出成型品質監控 方法，其包括有： ”第-階段之擷取模六壓力訊號’係利用模穴壓力感測 益配合程式擷取每一成型周期模穴内部壓力 料，並將其轉換且輸入電腦系統； “貝 第二階段之判斷成品品質’係將取得之模穴慶力訊號 貧料，入f品品質判斷系統，該系統即繪製模穴壓力曲線 ，運算其最大模穴壓力值及模穴壓力曲線積分值，以計算 出各種不同之模穴M力點’且利用曲線擬合技術配合多項 式為曲線擬合之配凑函數’緣製縱向及橫向原始擬合曲線 並刀別將其平行移動至X軸及γ軸方向之最大模穴壓力 值及模穴塵力曲線積分值位置的方向，以包絡每一、點資料 200424052 進而形成品質管制界限，而品質判斯計算 路模式ANN法則中之Sigm〇idal㈠，〇、用痛神經網 統判斷成品品質良劣之依據，將各模穴壓力田做亡： 壓力值及模六壓力曲線積分值分別代入品質管制 線方程式中，以計算各模穴壓力點之座落位置，厂之口 做為系統判斷成品品質良劣及其缺陷特徵之主要贫精據此’叮 所述之以模穴壓力為基礎之射出成型品質監 其監控方法開始前進行包括第一步驟之品質要㈣點及第 二=型參數之實驗設計的前置作業，經品質要求特點 ^中所產生之0口質缺陷，提出對該缺陷之參數實驗 並經實驗運算分析求得—組靈敏度較高之成型 二;且二再經變異數分析尋得貢獻度最高之兩項參數進 :幅度之調整，且開始將各種組合之成型參數進行訓 ^ ¥資料之取樣，即每組參數組合進行採樣50模，完 、後將每組所去樣之前25模進行資料訓練學習，後Μ模 進行該組品質測試，最後進行統計分析出誤判率最少之參 數、、且σ，並做確認實驗以確認該組成型參數之效果。 所述之以核穴麼力為基礎之射出成型品質監控方法， 金、經品質要求管制及提出參數實驗設計法求得最佳成型參 文後’於該成品品質判斷系統之成品品質判斷計算結束後 备即進行品質管制界限面積之計算，係將擬合曲二不同 松度旋轉’而找出在判斷準確率最佳下之面積最小為界定 乾圍’並經測試結果判斷成品良率是否符合標準，成品良 率未符合標準時，即提出改善方法，以調整成型表數至最 456 200424052 穩健狀態。 藉由上述進行監控前之前置作業、第一階段之擷取模 八C力5fi號及第二階段之判斷成品品質，可確實針對射出 成型品質之良劣進行監控，且於監控過程中，操作員只須 將模八C力δ孔號資料輸入成品品質判斷系統中，即可完成 成品監控作業，故可以最節省人力資源的方式，針對射出 f型品質之良劣進行監控，且該監控方法亦可對成品之缺 陷特徵，做更進一步的辨識，並達到極高的辨識率，以提 供操作員調整成型參數之依據，而有效提升射出成型後之 成品品質。 【實施方式】 為使貴審查委員可確實了解本發明之設計，及其它 發明目的與功效，以下兹舉出一具體實施例，並配合圖式 詳細說明如下： 有關本發明之具體實施設計，請參閱第一圖所示，本 t明以模穴壓力為基礎之射出成型品質監控方法，其主要 =進行監控前之前置作業（s丄）、第—階段之擷取模 八C力成唬（s 2 ) A第二階段之判斷成品品冑（s 3 ) 所構成。 進行監控前之前置作業（S1)係包括第一步驟之品 質要求特點（S1m第二步驟成型參數之實驗設計（ s 1 2 ) ’其中’第一步驟之品質要求特點（s 1丄）， 係針對於初步設定成型參數時，為避免成型品質產生太多 缺陷現象及祈望成型品質更為均_，而對成型品f提出之 4S? 8 200424052 要求，而第二步驟之參數實驗設計（s i 2 )，如 所示：對品質要求特點（S11)成品要求中所二 之品貝㈣，提㈣於錢陷之參數實驗設計 2^’經該實驗之運算與分析後求得—組靈敏度較高之 成型參數組合（S 1 2 2 ) ^ $ ° 4 ^ ^ π ，元成後再將該經變異數分析 後’寻付貝獻度取馬之兩組參數進行±5%幅度之調整（S 123) ’並開始將各種組合之成型參數進行訓練學習資 t之：樣J S t 2 4 )，亦即將每組參數組合進行採樣50 模’完成後將每組所去樣之前25模進行資料訓練學習， 後25模進行該組品質測試，於最後再進行統計分析出誤 判率最少之參數組合，並做確認實驗以確認該組成型參數 之效果。 第一階段之擷取模穴壓力訊號（S 2 )，如第三圖所 示包括有埋設模穴壓力感測器（s 2丄）、以程式擷取模 穴内壓變化之訊號（S 2 2 )及壓力訊號之轉換及儲存7 5 2 3 )，首先，在欲進行射出成型品質監控之模具公模 内埋設數組模穴壓力感測器·，當射出成型機將熔融塑膠= 入模穴内時’其埋設於公模内之壓力訊號感測器，即可夢 由sfl 5虎擷取程式來控制其開關，而開始擷取從炼膠充填到 填滿過程中，每一成型周期模穴内部之壓力變化的訊號資 料，並隨即將該擷取後之訊號資料輸入電腦系統，且令其 轉換成文子播’以及進行該資料之储存，而完成第一卩比^L 之擷取模穴壓力訊號（S 2 )。 本發明第二階段之判斷成品品質（S 3 )，其主要係 9 45§ 200424052 將轉換完成之模穴壓力訊號資料輸入成品品質判斷系統之 中’以進行相關程式運算而達成，如第四圖所示，該成品 口口貝判斷系統之程式流程依序為輸入模穴壓力資料（S 3 1 )、擬合曲線之繪製（s 3 2 )及成品品質判斷設計（ S 3 3 )，首先，在輸入模穴壓力資料（s 3丄）後，系 、、先即繪製出之模穴壓力曲線，如第五圖所示，並可利用程 式運算出該模穴壓力曲線之最大模穴壓力值及模穴壓力曲 線積刀值，且在系統不斷的重覆輸入及計算每一成型周期 =八壓力M化的訊號資料’而能獲得各種不同的模穴壓 (1〇)，以做為監控成品品質之主要判別參數。 田兀成k八壓力資料之輸入時，#予以進行擬 之繪製（S ^ 、 ’以計异品質管制區域之界限，而該擬 3曲線纟會製方、、表 '，主要係利用曲線擬合技術（Curve Fming)，並以你b -多項式微曲線擬合支配湊函數擬合點(Statistic Process control, SPO 3 monitors the position, nozzle pressure, or cavity pressure. Among them, due to melting: plastic mold cavity ㈣, the object it directly contacts is the mold wall itself. * Right, the surface of the branch wall The embedded signal sensor is the most direct and can be the most :: 'leather k solder glue flow state and pressure change in the branch cavity, and then the more accurate value can be used as a basis for adjusting parameters or monitoring product f, so ^ Quality control of mold cavity pressure is still the most appropriate method in terms of effectiveness 4S4 5 200424052 Quality monitoring method of cavity pressure 'Although it can be used for injection molding quality members: advance: monitoring', but the method is still in the process of monitoring It is necessary to monitor the quality of the finished product by operation and interruption. Therefore, it is impossible to stop people six times, / 5, and the defect characteristics of the finished product: ", waste ', and also special rough edges or short shots, etc. to make further & quot …, Method to effectively ensure the quality of the finished product after injection molding. [Inventive Content] = This is the 'quality monitoring method to improve the traditional mold cavity pressure of Lishu'. 'The inventor has developed a design, molding "Quality monitoring method", a good way to shoot and take the thirsty human resources, to monitor the quality of the molding quality, and to effectively identify the characteristics of the finished product, is its main purpose of invention., = 了Achieving the aforementioned object of the invention 'Technical hand used by the present invention; it is to provide a method for monitoring the quality of injection molding based on the force of the cavity of the mold, which includes: "The sixth phase of capturing the mold six pressure signal' system Use the cavity pressure sensing benefit program to capture the pressure material in the cavity of each molding cycle, convert it and enter it into the computer system; "Judging the quality of the finished product in the second stage" means that the cavity cavity strength signal will be poor. Material, into the product quality judgment system, the system draws the cavity pressure curve, calculates the maximum cavity pressure value and the cavity pressure curve integral value to calculate various M cavity points of the cavity, and uses curve fitting. The technical coordination polynomial is a fitting function for curve fitting. 'Make vertical and horizontal original fitting curves and move them in parallel to the maximum cavity pressure values and cavity dust in the X-axis and γ-axis directions. The direction of the position of the curve integral value is based on enveloping each and every point of data 200424052 to form the quality control limit, and Sigmoidal㈠ in the ANN rule of the quality judgment calculation mode. 0, the basis for judging the quality of the finished product using the pain neural network system , The pressure field of each cavity is killed: the pressure value and the integral value of the mold six pressure curve are substituted into the quality control line equation to calculate the position of the pressure point of each cavity, and the factory mouth is used as a system to judge the quality of the finished product. Based on this, the main poor essence of its defect characteristics is based on the injection molding quality based on the cavity pressure, and the monitoring method is performed before the monitoring method including the first step of the key points of quality and the second = type parameters. The preparatory operation of the model is based on the quality defects of 0. The parameters of the defect are proposed and analyzed through experimental calculation and analysis—the group with higher sensitivity is formed as the second one; The two parameters with the highest contribution are adjusted: the amplitude is adjusted, and the molding parameters of various combinations are trained ^ ¥ data sampling, that is, each group of parameter combinations is sampled 5 0 models. After completion, the 25 models before each sample are trained and learned, the M models are tested for the quality of the group, and the statistical analysis is performed to find the parameter with the least false positive rate, and σ, and a confirmation experiment to confirm The effect of this constitutive parameter. According to the method for monitoring the quality of injection molding based on the core cavity force, the quality of the finished product is judged in the finished product quality judgment system after gold and gold are controlled by quality requirements and the parameter experimental design method is obtained to obtain the best molding parameters. The backup is the calculation of the quality control boundary area, which is based on the rotation of the fitting curve 2 with different degrees of looseness to find the smallest area with the best judgment accuracy as the defined dry circumference, and the test results to determine whether the yield of the finished product meets Standard. When the finished product yield rate does not meet the standard, an improvement method is proposed to adjust the number of molding tables to a maximum of 456 200424052. Based on the above-mentioned pre-preparation operation before the monitoring, the first stage of the extraction mold eight C force 5fi number and the second stage of the judgment of the quality of the finished product, the quality of the injection molding quality can be accurately monitored, and during the monitoring process, The operator only needs to input the mold eight C force δ hole number data into the finished product quality judgment system to complete the finished product monitoring operation. Therefore, the most human-saving way can be monitored for the quality of the injected f-type quality, and the monitoring The method can also further identify the defect characteristics of the finished product and achieve a very high recognition rate to provide the basis for the operator to adjust the molding parameters and effectively improve the quality of the finished product after injection molding. [Embodiment] In order for your review committee to know the design of the present invention, as well as other objects and effects of the invention, a specific embodiment is given below, and it is explained in detail with the drawings as follows: For the specific implementation design of the present invention, please Referring to the first figure, this method is based on the mold cavity pressure based injection molding quality monitoring method, which mainly = the pre-operation before monitoring (s 丄), the first stage of the extraction mold C force formation (S 2) A is the second stage of the final product (s 3). The pre-operation before monitoring (S1) includes the quality requirement characteristics of the first step (S1m the experimental design of the molding parameters of the second step (s 1 2) 'wherein' the quality requirements characteristics of the first step (s 1 丄), In order to avoid too many defects in the molding quality and to hope that the molding quality is more uniform when setting the molding parameters initially, the 4S? 8 200424052 requirement for the molded product f is proposed, and the parameter experimental design of the second step ( si 2), as shown: for the quality requirement characteristics (S11) of the finished product requirements, the parameters of the product were improved from the money trap experimental design 2 ^ 'After the calculation and analysis of the experiment to obtain-group sensitivity The higher molding parameter combination (S 1 2 2) ^ $ ° 4 ^ ^ π, after Yuan Cheng analyzes the two sets of parameters after the analysis of the variation number, and the two parameters of the horse are adjusted by ± 5%. (S 123) 'and start to train the molding parameters of various combinations to learn the sample t: sample JS t 2 4), that is, to sample 50 sets of each set of parameter combinations' after completion of the 25 sets of samples before each set of samples Data training and learning, the last 25 models for this group of quality testing , To finally carry out a statistical analysis of a minimum rate of misjudgment combination of parameters, and experiments do confirm to confirm the effect of the constitutive parameters. The first phase of the cavity pressure signal (S 2), as shown in the third figure, includes a buried cavity pressure sensor (s 2 丄), and a program to capture the signal of the cavity pressure change (S 2 2). ) And the conversion and storage of pressure signals 7 5 2 3), first, the array cavity pressure sensor is embedded in the male mold of the mold to monitor the quality of the injection molding. When the injection molding machine melts the plastic = into the cavity 'The pressure signal sensor embedded in the male mold can dream to control its switch by the sfl 5 tiger capture program, and start to capture the inside of the mold cavity during the molding cycle from filling to filling. The signal data of the pressure change, and then the signal data after the capture is input into the computer system, and it is converted into a text sub-broadcast 'and the data is stored to complete the first cavity pressure of ^ L. Signal (S 2). The judgment of the quality of the finished product (S 3) in the second stage of the present invention is mainly based on 9 45 § 200424052. The converted cavity pressure signal data is input into the finished product quality judgment system 'to perform related program calculations, as shown in the fourth figure. As shown, the program flow of the finished product mouth and mouth judging system is sequentially inputting cavity pressure data (S 3 1), drawing of fitting curves (s 3 2), and final product quality judgment design (S 3 3). First, After entering the cavity pressure data (s 3 丄), draw the cavity pressure curve first, as shown in Figure 5, and use the program to calculate the maximum cavity pressure value of the cavity pressure curve. And mold cavity pressure curve product knife value, and the system repeatedly inputs and calculates each molding cycle = eight pressures of M signal signal data, and can obtain a variety of different cavity pressures (10) for monitoring The main judging parameters of finished product quality. When Tian Wucheng entered the pressure data of # 8, he shall draw a pseudo-drawing (S ^, 'to calculate the limits of different quality control areas, and the pseudo-3 curve will be used to make formulas, tables, etc.', mainly using curve fitting techniques (Curve Fming), and use your b-polynomial micro-curve fitting to dominate the fitting function to fit the points
(B) (C) y"a+bx "7 '-------L^J /=/ m „}-- «ΗΣχ,/ m /=/ m '^^--^ 117知)-匕)2 叫 /=/ 4S§ 10 200424052 如第六圖所示，該公式中之x、y分別為良品在各模 :壓力點（]_ 〇 )之最大模穴壓力值（pmax)及良品的模 穴壓力曲線積分值（Pindex )，在本發明之實施例中，係 針對各模穴壓力點（]L 〇 )進行曲線擬合，其（χ，y f 座標分別以各模穴壓力點（1 0 )之（Pmax，Pindex)座 標及（Pindex，Pmax)座標，帶入公式（A)進行計算， 配口 a式（b )、（ C )的計算，以獲得最趨向各模穴壓 力點（1 〇 )之縱向及橫向原始擬合曲線（2 1 )、（ 2 2 )，配合參閱第七圖所示，再分別計算位於χ軸方向最 :模穴壓力值（Pmax)之壓力點最大值（丄丄）與壓力點 取值（1 2 )，以及在γ轴方向模穴壓力曲線積分值（ Ρΐ*Χ)之壓力點最大值（"）與壓力點最小值（丄4 )，藉此，即可將原始擬合曲線（2丄）、（2 2 )，分 別平仃移動至w述X軸及γ軸方向之最大值及最小值（1 每」12)、（13)、（14)位置的方向，進而 ’。，點貝'料形成由四條曲線所圍成之品質管制界限（ ，且由於品質管制界限（3 0 )主要係利用每 型周期之模穴壓力變化的資枓所# # 用母成 的°代唬貝枓所形成’故該品質管制 _〇)，可確實做為判別成品品質之界定邊界。 管制：圖所示’當擬合曲線之繪製完成，並形成品質 制"限（3 0 )時，苴口併总生丨英 品區（… —貝官制界限（30)内即為良 並區八^ ) ’而品質管制界限（3〇)外則為劣品區， 。：刀為短射區（42)及毛邊區（43)，再 。 叩負判斷設計（s 3 3 )，該計算方式主要 °° 1 #々八王要係利用類神經 200424052 =路模式ANN法則中之si—卜卜"方法，以做 為糸統判斷成品品質良劣之依據，如良品、短射、毛邊等 "判斷方式係參考前述所完成之品質管制界限（3 〇 ) 主要利用程式再將各模穴壓力點（i 〇 )之最大模穴壓 f (Pmax)及模穴壓力曲線積分值（pindex)，分別代 質管制界限（3 〇 )之四條曲線方程式中，以計算各 =穴壓力點（i 〇 )之座落位置，亦即當成品在射出成型 過^中，如其模穴麼力點位於品質管制界限（3 〇 )内 2品區（4 1 )時，即該成品為良品，反之，當其模穴 :、:位於品ff制界限（3 0 )外之短射區（4 2 )及 σ 4 3 )時，則该成品為劣品，並可藉此判斷該成 口口之缺陷特徵為短射或毛邊。 月再 > 閱第四圖所示’本發明為使品質管制界限（3 $之準確性提升’當成品品質判斷設計（S 3 3 )結束 < ’即可繼續進行品質管制界限（30)面積之計曾，主 擬合曲線做不同角度之旋轉，以期找出在判麟確 瞀Γ二下之取小面積為界定範圍，當最佳之界定範圍計 二i ’即可實際運用於生產線上’進行品質監控測試 二：：結果來判斷成品良率是否符合標準，且將每 種：：組合取樣50模’之後再將各組曲樣之前25模作 為如丨練學習杳:1 ^ 取出於最測試㈣’最後再進行統計 ，成型^ 率最小之參數組合，完成後便將 ㈣成二ΪΓ于確認實驗’以確認該參數之可靠性，進而 多數組合之確認效果。又’如未在成品良率符合 4$1 12 200424052 標準時，即可再提出改善方法，係利用田口法改善成型灸 數’以調整成形條件至最穩健狀態（s 3 4)，例如：在 ' j U中叙ί見’其具毛邊之成品容易與良品相互混合， 代表叹疋之成型茶數不穩健因此，為改善成形良率，即利 用田口法（Taguchi Method)，配合如第九圖所示之Lu 直交表，計算出如第十圖所示之最佳參數組合表，且取樣 模之成品進行品質判斷測試，復以最佳參數組合表之 相關夢數值為依據’來改善成品品質並同時分離良品與毛 邊不良品在Pindex至Pmax圖的分佈位置，如此亦可使該 系統之判別正確率提高。 由上述各& ^又之貫施方式得知，本發明以模穴壓力為 基礎之射出成型品質監控方法之特點在於：藉由進行監控 纳之珀置作業（S 1 )、第一 ρ皆段之擷取模穴壓力訊號（ s 2 )及第二階段之判斷成品品質（s 3 )，即可確實針 ㈣出成型品質之良劣進行監控，且在監控的過程中，、操 作員只須將轉換完成之模穴壓力訊號資料輸人成品品質判 斷系統中’以進行相關程式運算，即可順利完成成品監控 作業’故在監控的過程中，操作員不須不斷的進行監控， 而能有效節省人力資源’且利用本發明之監控方法，亦可 對成品之缺陷特徵，如毛邊或短射等，做更進—步的辨識 ’且可達到極高的辨識率，以提供操作員調整成型參數之 主要依據’因此’可有效提升射出成型後之成品品質。 由以上的說明可知，本發明以模穴壓力為基礎之射出 成型品質監控方法，可以最節省人力資源的方式，針對射 4S2 13 200424052 出成型品質之M飞行監控， 徵’其確& 一相當優異之設計 【圖式簡單說明】 並有效辨識其成品之缺陷特 (一）圖式部分 你尽發明監控方法之流程圖 … q工 (eg 。 =二圖係本發明參數實驗設計之流程圖。 弟二圖係本發明擷 第四圖係本發明成。：：…號時之流程圖。 第― 質判斷系統之流程圖。 第五圖係本發明成品品 考圖。 、Μ断糸統計算模穴壓力點之參 統緣製擬合曲線時之參 統形成品質管制界限時 第六圖係本發明成品品質判斷系 考圖。 弟七圖係本發明成口 σ 只d城口口品質判斷系 之參考圖。 、 第八圖係本發明成品品質 算時之參考圖。蚜糸統進行成品品質判斷計 苐九圖係本發明利用田口 第十圖係本發明利用田口 (二）元件代表符號 (1 0 )模穴壓力點 (1 2 ) X軸壓力點最小值 (1 4 ) Y軸壓力點最小值 (2 2 )橫向原始擬合曲線 (4 1 )良品區 去時所採用之L 1 8直交圖。 斤°十异出最佳參數組合表。 (1 1 ) X軸壓力點最大值 (1 3 ) Y軸壓力點最大值 (2 1 )縱向原始擬合曲線 (3 〇 )品質管制界限 (4 2 )短射區 m 14 200424052 (4 3 )毛邊區(B) (C) y " a + bx " 7 '------- L ^ J / = / m „}-` `ΗΣχ, / m / = / m' ^^-^ 117 ) -Dagger) 2 Call / = / 4S§ 10 200424052 As shown in the sixth figure, x and y in the formula are the maximum cavity pressure values (pmax) of good products in each mode: pressure point (] _ 〇) The integral value of the cavity pressure curve (Pindex) of the good product and the good product. In the embodiment of the present invention, the curve fitting is performed for each cavity pressure point (] L 〇), and the (χ, yf coordinates are each cavity pressure The (Pmax, Pindex) coordinates and (Pindex, Pmax) coordinates of the point (1 0) are brought into the formula (A) for calculation, and the calculations of formulas (b) and (C) are assigned to obtain the most trending mold cavities. The vertical and horizontal original fitting curves (2 1) and (2 2) of the pressure point (10), as shown in the seventh figure, and then calculate the pressure in the direction of the χ axis: the pressure of the cavity pressure (Pmax). Point maximum value (丄 丄) and pressure point value (1 2), and the maximum value of pressure point (") and the minimum value of pressure point (丄 4) of the integral value of the pressure curve of the cavity in the direction of γ axis ), So that the original fitted curve (2 丄), 2 2), respectively, moving to the directions of the maximum and minimum values of the X-axis and γ-axis directions (1 each of 12), (13), and (14), and then '., 点 贝' material formation The quality control limit () surrounded by four curves (and because the quality control limit (30) is mainly used by the capital to change the mold cavity pressure of each type of cycle ## It is formed by using the mother's ° to replace the shellfish ' Therefore, the quality control _〇) can be used as a definitive boundary for judging the quality of the finished product. Control: As shown in the figure, when the drawing of the fitting curve is completed and the quality system " limit (3 0) is reached, the mouth and total Health 丨 British product area (...-within the official limit (30) is a good combination area ^ ^) 'and quality control limit (30) is a poor product area .: The knife is a short shot area (42) and The burr area (43), again. 叩 negative judgment design (s 3 3), the calculation method is mainly ° ° 1 # 々 八 王 wants to use the neuron-like 200424052 = road mode ANN rule of the si-bubu method, Use it as the basis for judging the quality of the finished product, such as good products, short shots, burrs, etc. " The judgment method is completed with reference to the foregoing The quality control limit (3 〇) mainly uses the program to calculate the maximum cavity pressure f (Pmax) and the cavity pressure curve integral value (pindex) of each cavity pressure point (i 〇), and substitutes the quality control limit (3 〇). In the four curve equations, the position of each = cavity pressure point (i 〇) is calculated, that is, when the finished product is in the injection molding process, if the force point of the mold cavity is within the quality control limit (3 〇), 2 products When the area (4 1), that is, the finished product is a good product, on the other hand, when the mold cavity :,: is located in the short shot area (4 2) and σ 4 3) outside the product limit (30), the finished product It is inferior, and it can be judged that the defect feature of the mouth is short shot or burr. Month > See the fourth figure, 'The present invention is to improve the quality control limit (3 $ accuracy improvement') When the finished product quality judgment design (S 3 3) ends < 'can continue the quality control limit (30) The area is calculated. The main fitting curve is rotated at different angles in order to find a small area under the limit of the determination. When the optimal range is calculated, two i 'can be used in the production line. Go to 'Perform quality control test 2 :: Results to determine whether the finished product yield rate meets the standard, and after each :: 50 samples in combination', then the first 25 molds of each group of music are used as a learning exercise: 1 ^ Take out at The most test ㈣ 'is finally counted and the parameter combination with the smallest ^ rate is formed. After completion, it will be converted into two ΪΓ in the confirmation experiment to confirm the reliability of the parameter, and then the confirmation effect of most combinations. When the yield rate meets the 4 $ 1 12 200424052 standard, an improvement method can be proposed. The Taguchi method is used to improve the number of forming moxibustions to adjust the forming conditions to the most stable state (s 3 4). For example, see 'J U 中 中 ί see'其 具 Its with The finished product of the edge is easy to mix with the good product, which means that the number of molding teas is not stable. Therefore, in order to improve the molding yield, the Taguchi Method is used in conjunction with the Lu orthogonal table shown in Figure 9 to calculate The best parameter combination table shown in the tenth figure, and the finished product of the sampling mold is subjected to quality judgment test, and based on the relevant dream value of the best parameter combination table as the basis to improve the quality of the finished product and separate the good and the burr defective products at Pindex. To the distribution position of the Pmax diagram, this can also improve the accuracy of the system. From the above-mentioned & consistent implementation methods, it is known that the injection molding quality monitoring method based on the cavity pressure of the present invention is characterized by : By performing the placement operation (S 1) of monitoring and receiving, the pressure signal of the cavity (s 2) for the first phase and the judgment of the quality of the finished product (s 3) in the second stage, the needle can be accurately pinched out. The quality of the molding quality is monitored, and during the monitoring process, the operator only needs to input the converted cavity pressure signal data into the finished product quality judgment system to perform related program calculations, that is, The finished product monitoring operation can be completed smoothly. “In the process of monitoring, the operator does not need to continuously monitor and can effectively save human resources.” Using the monitoring method of the present invention, the defect characteristics of the finished product, such as burrs or short Injection, etc., to make further-step identification 'and can reach a very high identification rate to provide the operator with the main basis for adjusting the molding parameters' so' can effectively improve the quality of the finished product after injection molding. From the above description, we can know that this Invented the injection molding quality monitoring method based on the cavity pressure, which can save the most manpower resources. For the M flight monitoring of the injection quality of 4S2 13 200424052, it has a very excellent design [simple drawing] Explanation] And effectively identify the defects of the finished product. (A) Schematic part of the flowchart of the monitoring method you try to invent ... q 工 (eg. = The second figure is a flowchart of the parameter experimental design of the present invention. The second picture is the present invention. The fourth picture is the present invention. :: No .: flowchart. No. ― Flow chart of quality judgment system. The fifth figure is a diagram of the finished product of the present invention. The parameters used in the calculation of the pressure points of the mold cavity when the system breaks the system are the parameters when the system fits the curve. When the quality control limit is formed, the sixth figure is a diagram for judging the quality of the finished product of the present invention. The seventh figure is a reference diagram of the quality judgment system for the Chengkou σ and Chengkoukou of the present invention. The eighth figure is a reference figure when calculating the quality of the finished product of the present invention. The aphid system is used to judge the quality of the finished product. Figure 9: The present invention uses Taguchi. Figure 10: The present invention uses Taguchi (II) element representative symbol (1 0). Cavity pressure point (1 2). Minimum value of X-axis pressure point (1). 4) The minimum value of the Y-axis pressure point (2 2), the transverse original fitting curve (4 1), the L 1 8 orthogonal plot used when the good product area is removed. The best parameter combination table for ten degrees. (1 1) Maximum value of X-axis pressure point (1 3) Maximum value of Y-axis pressure point (2 1) Longitudinal original fitting curve (3 〇) Quality control limit (4 2) Short shot area m 14 200424052 (4 3) Unedged area
464 15464 15