工業技術研究院

技術簡介

多聯式商用冷櫃廣泛使用於超市、量販店及冷鏈物流等場域，然而現行系統普遍欠缺協同控制機制，難以將壓縮機維持在最佳效率區間，導致能耗增加。為解決上述問題，本技術開發基於增強式深度學習的多聯式冷櫃協同控制技術，透過分析冷凍冷藏櫃之運轉資料，並以最低能耗、食品安全及製冷性能係數為目標，動態調整冷櫃設定溫度，藉由於多變環境下進行最佳化決策，可有效提升壓縮機使用效率並大幅改善整體節能效益。

Abstract

Multi-connected commercial freezers are widely used in supermarkets, hypermarkets, and cold chain. However, current systems generally lack collaborative control mechanisms, making it difficult to maintain the compressor within its optimal efficiency range, leading to increased energy consumption. To address these issues, this technology develops a multi-connected freezer collaborative control technology based on deep reinforcement learning. By analyzing the operating data of the freezers, and aiming at minimum energy consumption, food safety, and coefficient of performance, dynamically adjust the freezer’s set temperature. By making optimal decisions under varying environments, it can effectively improve compressor efficiency and significantly enhance overall energy savings.

技術規格

1.可應用於多聯式冷櫃交互影響時的複雜情境，透過即時掌握環境變化、多設備之運轉情況，自動生成最佳目標之設備運轉參數組合。 2.運用強化式深度學習演算法，對多聯式冷藏櫃，進行最佳化控制，可將壓縮機維持在最佳運轉狀態效率區間以最大化節能成效。

Technical Specification

1. Applicable to complex scenarios involving the interaction of multiple refrigerator units, it automatically generates the optimal combination of equipment operating parameters by real-time monitoring of environmental changes and the operation of multiple devices. 2. Utilizing deep reinforcement learning algorithms, it optimizes the control of multi-unit refrigerator units, maintaining the compressor within its optimal operating efficiency range to maximize energy savings.

技術特色

本系統以深度強化式學習為核心關鍵技術，打造新一代智慧能源管理系統，整合食品零售業場域中的多聯式冷櫃設備，並透過自動化的最佳化運轉控制策略，將壓縮機維持在最佳運轉狀態效率區間，同時兼顧節能效益、環境舒適性與商品品質穩定性。

應用範圍

人工智慧、能源管理系統

接受技術者具備基礎建議（設備）

無

接受技術者具備基礎建議（專業）

C＋＋ / C sharp / python程式能力

聯絡資訊