ارتقاء بازدهی پروژه حفاری: پیش ‌بینی و ارزیابی عملکرد ماشین رودهدر مبتنی بر هوش مصنوعی

رودهدر به عنوان یک حفار مکانیکی فضاهای زیرزمینی، اهمیت بالایی در پروژه‌های عمرانی و معدنی دارد و بررسی اقتصادی یک پروژه حفاری به شدت وابسته به پیش‌بینی عملکرد رودهدر است. نرخ خالص حفاری رودهدر یک معیار مهم در ارزیابی عملکرد آن به حساب می‌آید که از طریق روش‌های مختلفی می‌توان به تخمین آن پرداخت. در این تحقیق با استفاده از روش‌های هوش مصنوعی ماشین بردار پشتیبان (svm)، شبکه عصبی مصنوعی (ann)، نزدیک‌ترین همسایگی (knn)، جنگل تصادفی (rf) و رگرسیون خطی (lr)، مدل‌هایی به‌منظور پیش‌بینی عملکرد رودهدر ایجاد شد و دقت و مطلوبیت آن‌ها با هم مقایسه شد. متغیرهای ورودی در این مدل‌ها، شاخص کیفی سنگ (rqd) و مقادیر بازگشت چکش اشمیت r1، r2 و r3 می‌باشند و مدل‌سازی‌ها در نرم‌افزار داده کاوی orange انجام شده است. نتایج مدلسازی‌ها نشان می‌دهد که مدل‌ ماشین بردار پشتیبان با 0.847r^2= و 0.120rmse= از عملکرد قابل قبولی برای پیش‌بینی نرخ حفاری یا بعبارتی پیش‌بینی عملکرد ماشین رودهدر در حفاری تونل برخوردار است.

improvement of drilling project efficiency: ai-based roadheader performance prediction and evaluation

the economic assessment of a drilling project heavily relies on forecasting the drilling machine performance. the net cutting rate of a roadheader is a crucial parameter for performance evaluation, and it can be estimated using various methods. in this study, various models were developed to predict roadheader performance and their accuracy and desirability were compared. the models utilized artificial intelligence techniques such as support vector machine (svm), artificial neural network (ann), k-fold nearest neighbor (knn), random forest (rf), and linear regression (lr). the input variables for these models were rock quality designation (rqd) and the return values of the schmidt hammer r1, r2, and r3. the modeling task was conducted using the orange data mining software.introductionmechanized drilling in civil and mining projects is of paramount importance today, as it enhances productivity and efficiency in underground spaces and tunnels. various devices are employed for mechanized drilling, including impact hammers, tunnel boring machines (tbm), and roadheaders. roadheader machines offer several advantages over traditional methods and other devices. accurately, predicting roadheader performance and estimating key criteria like instantaneous cutting rate (icr) and operational cutting rate (ocr) is vital for project scheduling, cost estimation, and planning. several methods are employed for these calculations, but empirical models have limitations in understanding the complex data structure and relationships.methodology and approachesin this study, svm, ann, knn, rf, and lr modeling techniques were used, and their performance was compared. the dataset utilized pertain to data collected during the construction of the istanbul sewerage system tunnel. it includes four input variables: rqd and the return values of schmidt hammer r1, r2, and r3, as well as the output variable, the net cutting rate (q) of the roadheader machine. modeling was conducted using the orange software, which provides an intuitive and comprehensible platform for high-speed intelligent analysis.results and conclusionsthe results indicated that the svm model exhibited satisfactory accuracy and could be employed to estimate the net cutting rate (q) or predict roadheader performance. conversely, the k-fold nearest neighbor model displayed the weakest performance in estimation. moreover, to identify the influential variable in the prediction process, the importance of different variables was assessed. it was observed that rqd significantly contributed to the estimation of the target variable, followed by the importance of r1.