公豬採精自動化之電腦模擬

Abstract

傳統的手套法(Gloved-hand method)採精，有作業效率低、人員工作時間長，以及不符合人員作業安全與人體工學等缺點，而近來所發展的採精作業自動化應可解決上述的問題。目前台灣的農村普遍存在勞力短缺的問題，因此採精作業自動化是養豬產業發展的必然趨勢。為因應產業的發展，本研究擬建構一套電腦模式，模擬採精作業的現場操作，進行電腦模擬並找出關鍵影響因子以改善現況。 本研究依照試驗種豬場採精作業的現場配置分成五個時間區段：1. 豬欄移動至準備室，2. 準備室等待時間，3. 準備室移動至採精室，4. 採精時間，以及5. 採精結束由採精室回去豬欄的時間。首先在現場收集20頭公豬整個採精過程的時間數據，以此做為建立模式的基本資料，找出各區間所使用時間的平均值及標準差，作為模式中之設定參數。模擬程式的開發環境為Microsoft Visual Studio 2010，以C++撰寫，圖形顯示則使用OpenGL函式庫，程式可調整的變數有採精時間與豬隻行走速率，而使用者介面參數設定選項包含有採精豬隻數、採精室與準備室選用數、出豬順序等。 建立初步模擬程式後，再以不同試驗時間所收集的32頭公豬的採精時間數據與初步模擬程式的電腦模擬結果做比較，進行參數的調整，此修改後的程式則再與不同試驗時間所收集的37頭公豬實際數據做比較，用以驗證模擬程式。在模擬程式確立後，對於現有豬場設備的使用可以進行評估與規劃，比如測試變數可包含採精室與準備室開放間數、待採豬隻出欄順序以及趕豬員工數，經由重覆模擬執行後比較其試驗結果，找出最佳的規劃如下：1. 開放4個採精室可提升約43%之時間效率，2. 受限於採精室個數，增開準備室對於整體時間節省的效益不大，故不建議增開準備室，以及3.採用離準備室最近之豬欄先出豬，所需總時間為最少。

For semen collection, the traditional gloved-hand method had several disadvantages such as low working efficiency, intensive labors, not considering the safety of collectors and human factors. The newly developed automated semen collection technique should be able to solve the above problems. Recently, labor shortage is a common problem existed in rural area of Taiwan, and the automated semen collection is a must for the swine industry. To follow the development of industry, the aim of this study is to develop a model which is able to simulate the operation of semen collection, and computer simulations will be conducted to find out the key factors of improving performance. According to the semen collection operation of the testing breeding farm, the whole process were separated into five time slots in this study: 1. moving time from pig pen to the waiting room, 2. waiting time, 3. moving time from waiting room to semen collection room, 4. semen collection time, and 5. moving time from semen collection room back to pig pen. Firstly, 20 heads of boars were intensively sampling for individual time slots data during the whole collection process. The means and standard deviations for each time slot were found and used as the starting parameters for a draft model using those previously collected data. The computer program of the model was coded with C++ based on Microsoft Visual Studio 2010, and the OpenGL was used for graph displaying. The manipulated variables in this program include semen collection time and the walking speed of the boar, and the interface parameters need be set up before executing the program are: number of boars to be collected, numbers of semen collection room and the waiting room to be chosen for operation, and order of boars to be collected. The parameters of this draft model was adjusted after their simulation results being compared with data collected from another batch of 32 heads of boars. This modified model was further validated using the data collected from another batch of 37 heads of boars. This validated model could be used in evaluating and planning of the existing facilities of the breeding farm. For example, the tested variables could include the number of semen collection room and waiting room, the order of boars to be collected and the number of operator driving the boars. After conducting recursive simulations and comparing the tested results, the optimal planning was found as follows: 1. using 4 semen collection rooms simultaneously could save operation time up to 43%, 2. efficiency improving was not significant when increasing the number of waiting rooms owing to the limitation of number of semen collection room, and 3. the minimum total operation time was achieved if the nearest boar from waiting room was out first.