In-Process Monitoring Equipment in Heat Treatment Today
High volume production in heat treatment facilities today lends itself to continuous heat treatment techniques. Gear shafts, axles, steel coils, and many other products are heat treated in continuous pusher or roller hearth furnaces or semicontinuous low-pressure carburizing furnaces, where the products travel through the furnace in product baskets. Temperatures in the zones of these furnaces are carefully controlled to impart the correct properties to the finished product, and a quench and secondary heat treatment are often incorporated in the production line.
The most important elements of the heat treatment process are time and temperature. All heat treatment specifications require heating the workload to specific temperature levels for a specific time and cooling at a specific rate. Parameters vary depending on the shape and thickness of the part, and the type of alloy, etc.
Precise measurement of the product temperature is vital to ensuring the product meets the required heat treatment specification. It is difficult, if not impossible in some cases, to measure the product temperature in a continuous furnace due to the problems of trying to feed long thermocouples through furnace doors, atmosphere locks, water quench baths, etc. To effectively monitor the temperature of a product in a continuous furnace it is necessary to place the complete monitoring system in the product basket so that it can measure the product temperature as it travels with the product. An in-process monitoring system generally consists of a data logger programmed to record from up to ten thermocouples, which are placed in specified positions around the working area of the furnace. A thermal barrier is used to protect the data logger from the high temperatures and atmospheres in the furnace (Fig. 1). Temperature data is downloaded from the data logger after the system has passed through the furnace, and can be analyzed using specially developed software.
Coil annealing. Recrystallization annealing of carbon-steel and martensitic stainless steel cold-rolled strip coils generally is carried out in a roller-hearth furnace under a nitrogen atmosphere, with atmosphere locks at the exit and entrance of the furnace (Fig 2). It is important to achieve throughout the entire coil the specific temperature at which recrystallization occurs and hold for the time required.
Normally, the annealing temperature is controlled by the thermocouples placed along the heating zone of the furnace, which only measure atmosphere temperature above the coil. It is difficult to run long "trailing lead" thermocouples embedded in the coil to measure the workload temperature due to the atmosphere locks at the ends of the furnace. Use of an in-process monitoring system allows placing up to ten thermocouples within and around the coil (Fig.3), which provide information not only regarding the maximum process temperature achieved and the time above the specified temperature, but also data about the rate of heating, temperature uniformity across the furnace, hot and cold spots and rate of cooling.
Figure 4 shows data from a typical coil annealing cycle. Thermocouples embedded deep in the coil (number 2 and 5) reach similar maximum temperatures as all other thermocouples, but the time to reach the set temperature (780°C, or 1435°F) at both these probes can be up to 1 hour longer. The temperature-time curves from successive trials also can be used to predict and model the strip temperature, and thereby make the heat treatment process more effective.
Low pressure carburizing. Carburizing is more commonly carried out in sealed quench furnaces than in vacuum furnaces. However, today an increasing number of manufacturers use vacuum carburizing (low pressure vacuum carburizing) because of specific advantages it offers over more traditional methods.
Of primary importance in the process is to determine the precise time the component needs to spend at the specified temperature in the carburize and diffusion cycles to not only ensure that the components have the correct level and depth of carbon penetration, but also to optimize batch processing time. This requires attachment of thermocouples to the workload.
In single chamber (batch) vacuum furnaces, it is sometimes possible to attach thermocouples to the load by passing them through air tight seals in the furnace wall. In multichamber (semi-continuous) vacuum furnaces, the load moves automatically between sealed chambers making it impossible to feed thermocouples through the furnace. Passing a monitoring system through the cycle together with the workload allows measuring the precise load temperature (Fig. 5). Figure 6 shows the time at various parts of the cycle. This data can be used to reduce the time spent in the various zones while not sacrificing product quality by making small reductions to the time cycles while closely monitoring resulting product properties.
Annealing electrical steel laminations. Heat treating steel laminations for electric motors and transformers may have several specific operations including: burning off residual oil and relieving stresses from the cold working (stamping) operation; decarburizing the steel by annealing in an (typically) exothermic atmosphere; and blueing the steel (producing an oxide layer on the surface) by exposing the product to a super heated steam atmosphere during the cooling cycle. The heat treatment process must be tightly controlled to produce a finished product with the optimum magnetic properties (low core loss and high permeability).
For high production volumes, heat treatment of laminations is generally carried out in roller hearth furnaces with the products stacked inside trays or product baskets (Fig. 7). Furnaces generally have several zones, such as burn off, anneal/decarburize, controlled cooling, blueing and rapid cooling, which may be separated by atmosphere locks. The atmosphere locks and the fact that the product may progress at varying speeds through the different zones make it difficult to monitor product temperature throughout the process without the aid of an in-process monitoring system. Monitoring product temperature data helps to ensure that:
- Products reach the minimum burn off temperatures (370 to 480°C, or 700 to 895°F) for a specific time to remove process oils
- The coldest part of the charge within the product basket meets the critical time/temperature specification (e.g., 1-1.5 hours at 755°C, or 1390°F) and do not exceed a maximum specified temperature (e.g., 800°C, or 1470°F) in the annealing/decarburizing zone
- The controlled cooling rate in the first cooling zone is achieved (e.g., 755 down to 500°C, or 1390 to 930°F, over 1-1.5 hours)
- The correct time in the blueing zone is achieved
More companies are monitoring product temperature in every furnace on a regular basis as part of their quality control program to spot changes to process conditions, such as hot or cold spots, at an early stage, before product quality is affected. The data in the process discussed above (Fig. 8) indicates that the time at temperature specification was satisfied at only two thermocouple positions, that the maximum temperature was exceeded at one thermocouple position, and the rate of controlled cooling was too fast at all thermocouple positions.
Temperature uniformity surveys
Furnace surveying is another area that can benefit from monitoring from within the furnace. Temperature uniformity surveys are a requirement for companies operating in the aerospace industry and some parts of the automobile industry. In surveying a furnace, thermocouples are placed in the extremities of the furnace and, after a period of stabilization, the measured temperatures at these points are compared to the set temperature of the furnace. If all the measured temperatures have been within known tolerances for a specific time period, the set furnace temperature is raised to the next level and the operation is repeated. Surveys are generally carried out at set time periods (e.g., twice per year), or after any major repair work or maintenance has been carried out on the furnace.
Conducting temperature uniformity surveys is usually easier in batch furnaces than in continuous furnaces, but it can still be a problem. The furnace must be cooled down before setting thermocouples in the working area. The ends of the thermocouples are then led out through the furnace door where they are connected to a data logger to conduct the survey (Fig. 9).
The main problem for manufacturers, especially those with multiple furnaces and more than one plant, is downtime. An average survey can take up to 24 hours, and there is also the possibility of damage to refractory material as the furnace is repeatedly cooled and heated. However, the in-process monitoring system could be mounted in the product basket and charged and discharged as a normal load, thus eliminating the need to cool and reheat the furnace before and after the survey.
A downside to using in-process monitoring systems for uniformity surveys is not being able to see the survey in real time. Information could only be examined after the system was discharged from the furnace, and the data logger downloaded to a PC, which precluded observing the point at which all the measuring thermocouples converge within the tolerance band (Fig 10). This is important as the specified time at a survey level begins from this point. However, the introduction of RF transmission allows transmitting data from inside the furnace to a PC, and the entire survey managed from there.
Using in-process monitoring equipment for surveys requires a balance between size of thermal barrier and thermal duration. The barrier (Fig. 11) must keep the data logger at a safe working temperature throughout the survey; a larger barrier keeps the logger cool longer. However a thermal barrier with an overly large volume compared with the working volume of the furnace could affect the spread of the survey results.
The complexity of conducting a survey increases in a continuous furnace. If using normal methods, survey thermocouples must be fed into the furnace as the product moves from zone to zone. A disadvantage of this method is that no product can be fed behind the survey jig, resulting in lost production.
This problem is solved by placing an in-furnace monitoring system into a survey jig (Fig. 12), and loading this as you would a normal furnace load. Thus, product can be loaded in front of and behind the survey jig. The monitoring system can be discharged hot or through a water or gas quench, which saves valuable production time.
Potential savings in surveying
Recent tests show that an in-process monitoring system can cut the time for an average survey from 24 to 6 hours, which can mean considerable savings for medium to large operations. For example, a plant with an average of 10 furnaces per plant will conduct a minimum of 20 surveys per year. An average of 24 hours per survey translates to 480 hours of downtime per year. At $100 per hour furnace time, this equates to $48,000 per year, which could be reduced by 75% saving $36,000 per plant, per year. These figures refer to contract heat treatment plants, but in-house heat treatment operations can also realize big savings if surveying is a requirement.
Advancements in software
In the past, software packages supplied with monitoring systems were basic, mainly concentrating on screens to reset and download the data logger, plot time vs. temperature and show some basic calculations such as maximum temperature reached, etc. Newer software has multiple functions including:
- Full profile analysis including rate of heating and cooling, time at temperature, tolerance settings and file overlay
- Full function real-time analysis including alarms and RF signal condition indicator
- Full uniformity surveying functions including thermocouple and instrument correction factors, survey management with warnings of overshoot and full survey report compilation and printout