第一步：明确要测的指标
常规水质在线监测指标大致分几类：基础理化（pH、ORP、电导率、温度、溶解氧）、感官与悬浮类（浊度、悬浮物 SS、污泥浓度 MLSS）、有机物与营养盐（COD、氨氮、总氮、总磷）、消毒相关（余氯、二氧化氯、臭氧）、以及流量与液位。
不同指标的测量原理完全不同，也决定了后续的维护量与耗材成本。选型前先把「必须测」和「最好测」分开，预算有限时优先保证前者。
第二步：明确工况条件
- 介质：纯水、自来水、市政污水、工业废水、海水还是强腐蚀液体？介质决定电极材质、膜片材质与接液部件材质。
- 温度与压力：常温常压、高温（>80℃）还是带压管道？高温需要高温型电极或降温装置。
- 悬浮物与油脂：含油、含纤维、易结垢的介质要选带自清洗（机械刮刷、气吹、超声）或易拆洗的结构。
- 爆炸性环境：涉及危险区域时要先定区域划分，再定防爆型式。
第三步：确定安装方式
常见安装形式有浸入式（池内支架安装）、插入式（带球阀可带压插拔）、流通式（旁路流通池，便于拆装标定）、管道式（直接安装在工艺管路上）。
选型时常被忽略的一点：有没有安装位置和检修空间。浸入式需要池边固定点与起吊空间；插入式要考虑是否允许断流/带压安装；流通式需要取压点与排放点。
第四步：确定输出与通信
- 4-20mA：最通用，抗干扰好，二线制可同时供电，断线可通过 0mA 识别。
- RS485 / Modbus RTU：适合多点组网与远距离传输，便于接入 PLC 或上位机。
- 继电器输出：用于高低限报警或直接控制加药泵。
- 数字协议（HART、Profibus 等）：多用于高端过程仪表，可远程诊断与组态。
第五步：算全生命周期成本
在线仪表的总成本 = 仪表本体 + 安装附件（支架、护管、流通池、电缆）+ 耗材（电极、膜帽、试剂、泵管、标定液）+ 维护人工。
很多项目只比主机价格，结果后期试剂与电极的消耗远超预期。选型时建议把 1 至 2 年的耗材费用一起算进去。
第六步：确认可维护性
问清三件事：多久标定一次、标定需不需要标准液、耗材更换周期是多久。能远程诊断、耗材更换简单的仪表，长期看更省心。
常见选型误区
- 只看主机价格，忽略试剂与电极的长期消耗；
- 只给量程，不给介质成分与温度，导致电极材质选错；
- 忽略温度补偿，温度波动大的工况读数会漂；
- 信号电缆与变频器动力电缆同槽敷设，造成读数跳变；
- 没有预留检修空间，后期标定要搭脚手架。
把工况参数、介质成分、安装条件与输出要求列成一张表发给我们，我们可以按这套框架给出配置清单与选型建议。
Step 1: Define what to measure
Online water-quality parameters fall into a few groups: basic physical-chemical (pH, ORP, conductivity, temperature, dissolved oxygen), sensory and solids (turbidity, suspended solids, MLSS), organic and nutrient (COD, ammonia, total nitrogen, total phosphorus), disinfection (residual chlorine, chlorine dioxide, ozone), plus flow and level.
Different parameters use completely different measuring principles, which also determines maintenance effort and consumable cost. Split parameters into "must have" and "nice to have" before comparing prices.
Step 2: Define the operating conditions
- Medium: pure water, drinking water, municipal sewage, industrial wastewater, seawater or aggressive chemicals - this decides electrode, membrane and wetted-part materials.
- Temperature and pressure: ambient, high temperature (>80 C) or pressurised pipework - high temperature needs high-temperature electrodes or a cooling loop.
- Solids and grease: oily, fibrous or scaling media need self-cleaning (mechanical wiper, air blast, ultrasonic) or easily removable designs.
- Hazardous area: classify the zone first, then choose the protection type.
Step 3: Choose the installation form
Common forms are immersion (mounted on a bracket in the basin), insertion (with a ball valve for installation under pressure), flow-through (a bypass cell that is easy to remove for calibration) and in-line (directly in the process pipe).
The point most often overlooked is whether there is space and access for servicing. Immersion needs a fixing point and lifting clearance; insertion needs to know whether flow can be stopped; flow-through needs a sampling point and a drain.
Step 4: Define outputs and communication
- 4-20 mA: the most universal, good noise immunity, two-wire versions carry power as well, and 0 mA reveals a broken loop.
- RS485 / Modbus RTU: good for multi-drop networks and long distances, easy to connect to a PLC or SCADA.
- Relay outputs: for high/low alarms or direct dosing-pump control.
- Digital protocols (HART, Profibus): typical of higher-end process instruments, allowing remote diagnostics and configuration.
Step 5: Calculate the life-cycle cost
Total cost = instrument + installation accessories (brackets, guard tubes, flow cells, cables) + consumables (electrodes, membrane caps, reagents, pump tubing, buffer solutions) + maintenance labour.
Many projects compare only the instrument price and later find reagents and electrodes cost far more than expected. Include one to two years of consumables in the comparison.
Step 6: Confirm maintainability
Ask three questions: how often is calibration needed, does calibration require standard solutions, and what is the consumable replacement interval. Instruments with remote diagnostics and easy consumable change are cheaper to run.
Common mistakes
- Comparing instrument price only and ignoring long-term reagent and electrode cost.
- Giving only the range, not the medium composition and temperature, so the electrode material is wrong.
- Ignoring temperature compensation, so readings drift when temperature varies.
- Running signal cable in the same tray as variable-frequency drive power cable, causing jumping readings.
- Leaving no service space, so calibration later needs scaffolding.
Send us your operating parameters, medium composition, installation conditions and output requirements and we will return a configuration list based on this framework.